Engineering in Chalk

Lawrence, Preene, Lawrence and Buckley

ISBN 978-0-7277-6407-2
https://doi.org/10.1680/eiccf.64072.663
ICE Publishing: All rights reserved

Chalk fractures geometry: a comprehensive description

of fracture surfaces
M-L. Wattier*1, F. Descamps1, S. Vandycke1 and J-P. Tshibangu1
1
UMONS, Mons, Belgium
*
Corresponding Author

ABSTRACT Beyond the fundamental identification and characterisation of faults and joints in chalk, the fine description of fracture per-
meability, by means of analysing roughness, preferential flow channels and related tortuosity, is crucial in multiple applications such as un-
derground stability studies, tunnelling, oil and gas production and related water injection, pollution control around contaminated zones, etc.
Chalk exhibits a plastic behaviour beyond the elastic limit. The shearing, potentially extensive pattern of failure, leads to frequent and
abundant fractures. As dynamic plate tectonics give form to various tectonic contexts, the resulting fractures can provide important flow
channels within the porous media. Fracturing and alteration processes induce specific surfaces. Structures such as hackles, striations, or
plumose patterns may emerge. Fault surfaces can display ridges or deep linear channels affecting the chalk matrix. The understanding of
fractures surface geometry by observation and qualification is hence crucial, as it directly affects fluid transport. Textural modifications and
surface properties (such as fracture plane roughness) derive directly from the fracturing mechanism and the flow process. As a result, the
description and classification of rock fractures can be conducted via systemic cause-to-effect approach. In this study, we use a cross-motion
table with CCD laser sensor to scan natural fractures in chalk as well as lab generated fractures. Working with samples collected in exten-
sions of the North West European basin, the 3D characteristics of the surfaces are put in context with the tectonic background. Surface
measurements will be compared to conventional techniques such as the JRC and other current methods (fractal and statistical fracture de-
scription), in view to consider an extension of fracture characterisation methods in three dimensions. Establishing a semi-automated meth-
od, this study will enable visualisation and classification of natural joints and faults, in view to improve fracture permeability depiction in
fundamental studies or practical applications.

1 INTRODUCTION
Artificial fractures are commonly referred to in order
to characterize surface roughness state (Nemoto et al.
2009), to develop stability models or permeability
models. Studies are generally done on crystalline
rocks, such as diorite or granite samples (Vogler et
al, 2017).
Roughness can be characterized by visual observa-
tion of fractures and in some cases, the JRC (Barton
& Choubey 1977) values are issued from practical
measurements on profiles using gauges and compar-
ing with Barton’s reference roughness profiles. Many
authors characterize fracture surface
itfr.63938.001.3d 3
with optical or
itfr.63938.001.3d 3
samples are seldom put into their geological and tec-
tonic context.
In this work we propose to cross-reference fracture
surface descriptions by combining visual observa-
tion, JRC profiles, and roughness characterization via
statistical and fractal indices issued from the pro-
cessing of surface laser scans. The observation of a
set of natural chalk samples displaying joints or
faults, is accompanied with - when available, detailed
orientation within the local tectonic structural con-
text. Artificial fractures generated in the same chalk
material by means of tensile or shear tests are charac-
terized with the same set of criteria as to provide fair
comparison.
04/02/2018 05:14:20pm 04/02/2018 05:14:20pm

laser techniques where precision generally ranges
from a few to several x10µm. Data can be processed
to deliver calculated JRC, fractal dimension or statis-
tical parameters. In many cases, only some of those
techniques are used in a single study, and the rock
Chalk fractures samples from locations in Harmi-
gnies (Belgium) and Cap Bland-Nez (France, Bou-
lonnais region) are here presented (Figure 1, Faÿ-
Gomord 2017).

663
Engineering in Chalk

Figure 2. Example of plumose and ribs on a joint in Obourg chalk

Formation Harmignies Belgium. From Vandycke 2002.

Figure 1. Location of chalk deposits in the North-West European

basin. The chalk of Harmignies in the Belgian Mons basin and the Table 1. Natural sample description and location, with main tec-
Blanc-Nez Chalk Formation from the French Boulonnais region (1 tonic systems’ direction. MJ: master joint, J: joint, Js: stratigraphic
and 2 respectively). Modified from Faÿ-Gomord 2017. joint, Fn: normal fault, Fd: dextral fault.
Sample Mode Struc- Strike dip, dip Scans
The Upper Cretaceous Campanian Obourg Chalk origin ture (striation)
Formation taken from Harmignies presents a finely Harmignies I MJ N20°, 87°E 21
grained lithology, is of very white colour, without
II Fn N121°, 70°S 6
chert. The clay-rich blue chalk of the Blanc-Nez
I Js / 5
Formation was collected at Petit Blanc-Nez and
I MJ N130°, 88°S 5
Grand Blanc-Nez sites. It comes from the base of the
Upper Cretaceous Series (Late and Middle Cenoma- / Js / 5
nian, in each respectively). I J N106°, 85°N 20
II Fn 120°N / ESE, 80°S 5
2 SAMPLE COLLECTION Blanc-Nez II Fn N65°, 60°S (85°N) 8
I J N105°, 35°S 5
Rock samples collected from outcrops were orientat- I or II J N-S or E-W or N135° 5
ed, with layer direction and dip measured on site. II Fd N25°, 75°N (10°N) 5
For fault surfaces, the direction of striation was also II Fd N25°, 75°N (10°N) 5
measured. The natural samples selected for this study
were limited to those with little to no weathering vis- 3 LABORATORY WORK
ible, to allow for better comparison with artificially
created fractures. Chalk blocks were identified and Structural features on specific samples were observed
collected for their relevance in terms of lithological in detail before and during each step of the process.
representativeness, suitable orientation with regards Compression, tensile and shear tests were conducted
to the major tectonic systems for their area, and for to characterize the mechanical strength of each type
bearing representative geological features found on of lithology studied, and produce artificially fractured
joints and faults in the tectonic structures of interest samples for further comparison with natural joints
(i.e. striation, plumose, ribs, twist hackles – Figure 2, and faults. Natural and lab generated fractures were
Vandycke 2002). The most relevant blocks were set then scanned by a high resolution laser to capture
aside for laser scanning of the natural geological fea- surface geometry. Additionally an in-house built
tures. The remaining blocks were cored to make cyl- MATLAB script calculated roughness indices.
inders for lab testing.

664
 Wattier, Descamps, Vandycke and Tshibangu
Wattier, Descamps, Vandycke and Tshibangu
Wattier,
Wattier,
Wattier,
Wattier, Descamps,
Descamps,
Descamps,
Descamps, Vandycke
Vandycke
Vandycke
Vandycke and
and Tshibangu
Tshibangu
and Tshibangu
Wattier, Descamps, Vandycke andand Tshibangu
Tshibangu

a. Indirect tensile test b. 45° shear c. 30° shear

a. Indirect tensile test b. 45° shear c. 30° shear
a.
Figure
a. a.
a. Indirect
Indirect
Indirect
3.
Indirect tensile
tensile
Blanc-Neztensile
tensile test test
testtest
Formation samplesb.
b. b.
45°
b.
45°: 45°
a.
shear shear
shear
45° shear
Brazil test. The surface is planar c.
c.30°c. 30°
30°
c.planar 30°
with
shearshear
shear
shear
a very homogeneous, granular texture. Geometry
a. Figure
Indirect3.tensile
Blanc-Neztest Formation samples b. 45° shear: a. Brazil test. The surfacec.is30° shearwith a very homogeneous, granular texture. Geometry
category
Figure ‘2-1’.
3. b. The surface displays striation
samples a.in the direction of shear, marked by a green line.homogeneous,
The surface isgranular
somewhat planar. Ge-
category
Figure
Figure
Figure
Figure
ometry 3.3. 3. Blanc-Nez
‘2-1’. b. The
Blanc-Nez
Blanc-Nez
Blanc-Nez
3. category
Blanc-Nez ‘2-3’.
Formation
surface
Formation
Formation
Formation
c.
displays
Formation samples
samples
samples
Plumose-like a.a.::Brazil
striation
a.: :Brazil
samples
:geometry a.inBrazil
Brazil the
Brazil
test.
is test.
visible
test.
direction
test.
The The
test.
The The
The
surface
on the
ofsurface
shear,
surface
surface
surfaceis planar
surface,
is
marked
isisplanar
is planar
planar
planar
with
which
by
withwith
a with
with
is very
aa very
aaagreen
very
very line.
very The surface
homogeneous,
homogeneous,
homogeneous,
homogeneous,
relatively rough and
isgranular
somewhat
granular
granular
granular
very texture.
uneven.
texture.
texture.
texture. Geometry
planar.
texture. Ge-
Geometry
Geometry
Geometry
Geometry
Geometry cate-
category
ometry
category
category
category ‘2-1’.
category
‘2-1’.
‘2-1’.
‘2-1’. b. Theb.
‘2-3’.
b. The
b. Thec.
The surface
surface
surface displays
Plumose-like
surface displays
displays
displays striation
geometry
striation
striation
striation is
in
in thein
the
in the
visible direction
on
direction
the the
direction
direction of of shear,
surface,
shear,
of
of shear,shear, marked
which
marked
markedmarked is
by bya
by a green
relatively
green
a green line.
rough
line. The
line. The
and surface
very
surface
The surface is is
uneven. somewhat
somewhat
is Geometry
somewhat planar.
cate-
planar. Ge-
planar. Ge-
Ge-
category
gory ‘2-1’.
‘2-2’. b. The surface displays striation in the direction of shear, marked by by a green
a green line.line.
TheThe surface
surface is somewhat
is somewhat planar.
planar. Ge-Ge-
goryometry
ometry‘2-2’.
ometry
ometry category
category
category
category ‘2-3’.
‘2-3’.
‘2-3’. c.
‘2-3’. c.
c. Plumose-like
Plumose-like
Plumose-like
c. Plumose-like geometry
geometry
geometry
geometry is is
is
is visiblevisible
visible on
visible on
the
on the
the surface,
surface,
surface, which
which
which is is relatively
relatively
is relatively rough
rough
rough and
and very
very
and very uneven.
uneven.
uneven. Geometry
Geometry
Geometry cate-
cate-
cate-
ometry category ‘2-3’. c. Plumose-like geometry is visible on on
thethe surface,
surface, which which is relatively
is relatively rough roughandand
veryvery uneven.
uneven. Geometry
Geometry cate-
cate-
gorygory
gorygory
‘2-2’.‘2-2’.
‘2-2’.
‘2-2’.
gory ‘2-2’.
a. Indirect tensile test b. 45° shear c. 30° shear d. 60° shear
a. Indirect tensile test b. 45° shear c. 30° shear d. 60° shear
a. a. Indirect
Indirect
a. Indirect tensile
tensile test
tensile test
test b.b. b.
45°
b.
45° 45°
45° shear
shear
shear shear c.
c. c.
30°
30°c. 30°
30°
shearshear
shear
shear d.
d. 60°
d.d.60°
60° shear
shear
60°
shearshear
a. Indirect
a. Indirect tensile
tensile testtest b. 45° shear c. 30° shear d. 60° shear

Figure 4. Obourg Chalk Formation samples : a. split by Brazil test. The middle of the sample was lost, it had typically split in several
Figure 4. Obourg Chalk Formation samples : a. split by Brazil test. The middle of the sample was lost, it had typically split in several
plane-like,
Figure brittle
4. sheets. Scales-like patterns are seen, a. withbysmooth surfaces, and zones where the rock seems itittoto have been chipped off.
plane-like,
Figure
Figure
Figure
Figure
Geometry4. Obourg 4. Obourg
brittle
4.4.category
Obourg
Obourg
Obourg Chalk
sheets.
Chalk
Chalk
Chalk
Chalk
Formation
Scales-like
Formation
Formation
Formation
Formation
‘2-2’. b.
samples
patterns
samples
samples
samples
samples
Striations cover a.a.::seen,
: a.:are
:split
the a.bysplit
split
split with
split
surfaceby byBrazil
Brazil by Brazil
smooth
Brazil
Brazil test.
test.
test.test.
consistentlyThe The
test.
The The
surfaces,
The
middle
along
middle
middleand
middle
middlezones
of
of the
the shear
of
thethe
ofthe
of the
sample
sample
where
sample the
sample
samplewas
direction,
was
was was
rock
was
lost,lost,
marked
lost,
itseems
lost, ithad
lost,
itby
had had had
had
typically
a green
typically
have
typicallybeen
typically
typically split
split
line. in
split
chipped
split
split
The
in
ininsurface
severalin several
off.
several
severalseveral
plane-like,
Geometry
plane-like,
plane-like,
plane-like, brittle
category
brittle
brittle
brittle sheets.
‘2-2’.
sheets.
sheets.
sheets. b.Scales-like
Striations
Scales-like
Scales-like
Scales-like patterns
cover
patterns
patterns
patterns the
are
are are
are
seen,seen,
surface
seen,seen,
with with
withwith smooth
consistently
smooth
smooth
smooth surfaces,
along
surfaces,
surfaces,
surfaces, the
and
and and
shearzones
zones
and
zones zones where
direction,
where
wherewherethe
the the rock
marked
rock
the
rockrock seems
by
seems
seems seems a
toto to
green have
have
to
have been
line.
been
have
been The
been chipped
chippedsurface
chipped
chipped off.off.
off. off.
plane-like,
plane brittle sheets.
is somewhat Scales-like
undulating. patternscategory
Geometry are seen,‘2-3’. with smooth
c. The surfaces,
surface is and
very zones
uneven, where the rock
broken seems toishave
up. Striation been chipped
however still visibleoff.on
Geometry
plane
Geometryis
Geometry
Geometry somewhat category
category
category
category ‘2-2’.
undulating.
‘2-2’. b.
‘2-2’.
‘2-2’. b. b.
b. Striations
Geometry
Striations
Striations
Striations covercover
category
cover
cover the
the the
the surface
‘2-3’.
surface
surface
surface c. consistently
The surface
consistently
consistently
consistently is along
along
along very
along the
the the
uneven,shear
shear
the
shear shear direction,
broken
direction,up.
direction,
direction, marked
Striation
marked
marked
marked by
by by
is
by
a a aa green
however
green
green line.
still
line.
green
line. The
line.
The The
visible
The surface
surfaceon
surface
surface
Geometry
some category
areas. Geometry b. Striations
‘2-2’.category ‘2-5’. cover
d. The thesurface
surfaceis consistently
highlyc. broken along
up. Duethe shear
to the direction,
higherbroken marked
lateral forcebycomponent,
a green line.parts Theofsurface
thevisible
rock on
someplaneareas.is
is somewhat
Geometry undulating.
category Geometry
‘2-5’. d. The category
surface is‘2-3’.
highly The
brokensurface is very uneven, up.up. Striation is
is however still
plane
plane
plane
was isissomewhat
isplane somewhat
somewhat
crushed somewhat
under
undulating.
undulating.
undulating.
undulating.
confinement.
Geometry
Geometry
Geometry
Geometry
Geometry
category
category
category
category
category ‘2-3’. ‘2-3’.
‘2-3’.
c.
‘2-5’. c.c.The
‘2-3’.
The The
c. The isup.
surface
surface
surface very Due
isisvery
surface very to
very the
uneven,
isuneven, higher
uneven,
uneven, broken
broken lateral
broken
broken up.up. force
Striation
up.
Striation component,
Striation
Striation isishowever
is however however parts
howeverstill
still of
still thevisible
visible
still
visible
visible onrockon on
on
was some
some crushed
some areas.
areas.
areas. Geometry
under
Geometry
Geometry category
confinement.
category
category ‘2-5’.
Geometry
‘2-5’. d.
‘2-5’. d.
The
d. The
category
The surface
surface ‘2-5’.
surface is is highly
highly
is highly broken
broken
broken up. up.Due
up. Dueto
Due to
the
to the higher
higher
the higher lateral
lateral force
force
lateral force component,
component,
component, parts
parts
partsof of
thethe rock
rock
some
some areas.areas. Geometry
Geometry category
category ‘2-5’.
‘2-5’. d. The
d. The surface
surface is highly
is highly broken
broken up.up.
DueDue to the
to the higher
higher lateral
lateral forceforce component,
component, parts
parts of the rockrockrock
of
of the the
waswas
was crushed
crushed
crushed under
under
under confinement.
confinement.
confinement. Geometry
Geometry
Geometry category
category
category ‘2-5’.
‘2-5’.
‘2-5’.
waswas crushed
crushed underunder confinement.
confinement. Geometry
Geometry category
category ‘2-5’.
‘2-5’.

a.
a.
a.
a. a.a. a.
Figure 5. Number of 30 mm x 30 mm scans (MJ: Master Joint, J: 172µm
Figure 5. Number of 30 mm x 30 mm scans (MJ: Master Joint, J:

YYY
Joint, F: Fault) for eachofchalk typexstudied. The simplified litholo- Y 172µm
Figure
Joint, 5. Number 30 mm 30 mm scans (MJ: Master Joint, J: J:
Figure
Figure
gy 5.F:5.
Figure
Figure
column
Fault)
5. Number
Number
shows
for
5.Number
Numbereach
ofof30
ofObourg
30 30
of
mm chalk
mm
30
mm mmxtype
xChalk
30 x30
30
mm studied.
mmmm
scans The
scans
xFormation
30 mm scans
scans
(MJ:(MJ:simplified
(MJ:
(MJ:
Master
from
Master
Master
Master litholo-
Joint,
Harmignies
Joint,
J:Joint,
Joint, inJ: J:
Y 172µm
172µm
172µm 172µm
172µm
Joint, F: Fault) for each chalk type studied. The simplified litholo- Y

Y
gy column
Joint, F:
Joint,
Joint, shows
Fault)
F: for for
Fault)
F: Fault) for Obourg
each
for
each each
chalkChalk
chalk
chalktype
type Formation
studied.
type studied.
studied.
TheThe from
The The Harmignies
simplified
simplified
simplified litholo-in
litholo-
litholo- Y Y

XX
Joint,
the F: Fault)
Mons Basin each
and chalk
the type
Boulonnais studied.
Blanc-Nez simplified
Chalk litholo-
Formation’s Y Y 172µm
172µm
30mm

gygy
the
gy Mons
gy column
column Basin
column
column shows
showsshows theObourg
andObourg
shows Obourg Boulonnais
Obourg Chalk
Chalk
ChalkChalk Formation
Blanc-Nez
Formation
Formation
Formation from
Chalk
from
from from Harmignies
Formation’s
Harmignies
Harmignies
Harmignies inin in
in
30mm

gyposition
column shows
withinBasinObourg
the Cretaceous. Chalk Formation from Harmignies in 172µm
the Mons and the Boulonnais Blanc-Nez Chalk Formation’s 172µm

XX
position
the
thethe Mons
the
Mons within
MonsBasin
Basin the
Basin
andandandCretaceous.
and
thethe the Boulonnais
the Boulonnais
Boulonnais Blanc-Nez
Blanc-Nez
Blanc-Nez Chalk
Chalk Formation’s
Chalk Formation’s
Formation’s X 172µm
172µm
172µm
30mm
30mm

Mons Basin Boulonnais Blanc-Nez Chalk Formation’s

30mm

X
30mm
30mm

position
position
position
position within
within
within
within thethe Cretaceous.
Cretaceous.
the Cretaceous.
position within thethe Cretaceous.
Cretaceous. X
X XX X

b. 30mm
b. 30mm
b. b.
b.
b. b. 30mm30mm30mm
30mm
30mm
Figure 6. a. Example of scan on Obourg Chalk For-
Figure 6. a. Example of scan on Obourg Chalk For-
mation sample.
Figure 6. a.b.Example
Schematics
of of a on
scan scanning
Obourgsequence
Chalk For-
mation
Figure
Figure 6. sample.
Figure
Figure6.6.Example
a. 6.
a. a. b.Example
Schematics
a. Example
Example ofscan
of of
scanscan
of of
onaObourg
onscan
on scanning
Obourg
on
Obourg Obourg sequence
Chalk
Chalk
Chalk
Chalk For-
For-
For-For-
with dimensions.
mation sample. b. Schematics of aa scanning sequence
with
mationdimensions.
mation
mation sample.
sample.
sample. b.
b. Schematics
b. Schematics
Schematics of
of a a
of scanning sequence
scanning
scanning
mation sample. b. Schematics of a scanning sequence sequence
sequence
withwith
with
with dimensions.
dimensions.
dimensions.
dimensions.
with dimensions.
663
663
665
663
663663
663 663
Engineering in Chalk

3.1 UCS tests tors on air cushion provide precise linear lateral dis-
placements along the X and Y perpendicular direc-
Cylinders were cored perpendicularly to sedimenta-
tions. Precision was set to 172µm along X and Y, and
tion. Tests were performed according to ISRM stand-
+/-30 µm over the Z readings. While this does not
ards (Fairhurst & Hudson 1999) on a stiff frame.
allow capture of features at grain scale, it was suffi-
Blanc-Nez Chalk Formation revealed to be more
cient to identify small surface variations and general
competent than Obourg Chalk Formation.
surface geometry (waviness).
Table 2. Average unconfined compressive strength over 5 samples The scanning sample area was set to 30 mm x 30
for each site. The average UCS for Blanc-Nez chalk is clearly
higher than for the Obourg chalk from Harmignies. mm as to use a representative (undamaged) area on
the lab generated fractures. Roughness indices being
Location Age Av. UCS Std dev.
[MPa] [MPa] sensitive to scale effects, the same dimensions (30
Harmignies Campanian 5.5 0.7 mm x 30 mm) were used for natural surface scans.
Blanc-Nez Cenom. Inf. 19.1 4.2
Altogether, 143 scans were performed (Figure 11).

3.5 Roughness indices: Ra, σa, Z2, Dvar and Dyard

3.2 Shear tests
Shear tests were conducted on cylinders of 40 mm
length by 40 mm diameter, using a set of two trun-
cated cylindrical metal cradles to hold the samples at
a set angle of 45°. Additional support bases allow for
extra tilt, so that the load during the test could also be
applied at 30° or 60° (Figure 7), allowing different
normal stresses on the shearing plane.
The resulting ASCII data was processed by a
MATLAB script. The calculated indices were: the
average asperities height Ra with its standard devia-
tion σa, the asperities heights RMS deviation Z2, and
the fractal dimensions for the yardstick method Dyard
and the semivariogram Dvar.
Ra (and related σa) and Z2 have been widely used
in the literature as statistical parameters to describe
rough fracture surfaces. Fractal dimensions were cal-
culated from the yardstick method and the semivari-
ogram method (Brown, 1987), (Ge et al. 2013,
Develi & Babadagli, 1998). The indices Ra and σa are
taken with regards to the median of the Z values
measured.

3.6 JRC determination

Figure 7. Shearing frame and steel cradles to set up shear plane at
45°, 30° or 60°.
3.3 Brazilian indirect tensile tests
Indirect tensile tests were conducted following the
Brazil test (ISRM 1978). Due to expected low
strength of the chalk samples, the cylinders were
fractured on a small compression frame under manu-
al loading.
3.4 Laser scanning
The Keyence LK-081 laser scanner provided asperi-
ties height readings over 3D surfaces. It is mounted
on a stiff cross-motion frame, where two stepper mo-
The JRC profiles were drawn using standard steel
gauges. Lines were traced along two perpendicular
directions, corresponding to the X and Y axes of the
scans. The natural rock samples being larger and of-
fering a greater range of features, at least two lines of
up to 15cm were taken in each direction.
4 SCANNED SURFACE CATEGORIZATION
Upon observation of the natural and lab generated
surfaces, it was clear that their analysis would require
grouping scanned surfaces in categories to ease com-
parison. The JRC profiles were of limited use in the-
se circumstances as the chalk sometimes presents
smooth but undulating or jagged surfaces, or, in the
contrary, rough but pseudo-planar surfaces.

666
 Wattier, Descamps, Vandycke and Tshibangu

An original convention is suggested: a two-digit 5.2 JRC profiles

code describing granularity and general surface ge-
The initial comparison of JRC profiles between natu-
ometry. The first digit, from 1 to 3, qualifies the tex-
ral and lab generated fractures is here inconclusive,
ture of the rock surface, with 1 corresponding to a
mostly as the parameter is poorly suited to such small
smooth texture and 3 to a very granular texture. The
samples (30 mm x 30 mm).
second digit, varying from 1 to 5, describes the com-
plexity of the surface geometry, with a value of 1 for
planar fracture and 5 for a very heavily broken-up
surface. Refer to Figure 3 & 4 for illustration. For
example a surface ‘Geometry category 2-5’ (slightly
grainy texture, highly wavy surface) was common on
the base of natural twist hackles.

5 RESULTS
Figure 8. Examples of JRC profiles of lab generated fractures
Blanc-Nez samples bn3 and bn5, both visually categorized ‘2-2’:
5.1 Visual observation the linear profiles do not fairly reflect the general surface state.

5.3 Roughness indices

5.1.1 Lab generated fractures
Preliminary analysis of roughness indices shows that
Cylindrical samples fractured in the lab display a while Ra is not conclusive, Z2, Dvar and Dyard display
wide range of surface geometries and textures (see increasing values (roughness) with increasingly com-
examples Figures 3 & 4). Striation is generally visi- plex geometries. A few samples however skew those
ble on sheared samples, but the amplitude of the se- general trends; the P10 sample in particular, a fault
cond order roughness (waviness) is variable depend- plane in the Blanc-Nez which displays high rough-
ing on the set shear plane. ness values. It was removed from the dataset of Fig-
Surfaces issued from the Brazilian method tend to ure 10 to better highlight general trends.
be pseudo planar with a rough, grainy texture, while As expected, natural fractures have smaller aver-
surfaces issued from shear tests show some degree of age index values (smoother surfaces), than the lab
striation and increased jaggedness. At 45°, shear sur- generated ones which also display greater standard
faces are pseudo planar with clear striation. As the deviation around their average value.
shear angle is set to 30°, i.e. when the shear direction
tends to approach the cylinder’s axis and the axial
load is greater than its lateral component, fracture
planes display ribs and undulations. At 60° shear an-
gle, surfaces have a most irregular appearance, no
striation visible anymore. Small zones were observed
over the samples where the rock had been locally
crushed under confinement.

5.1.2 Natural fractures

Natural fractures, in a general way, are less rugose
than the fresh, lab generated fractures. Master joints
and joints tend to display plumoses on centimetric to
decimetric scale. Ribs and twist hackles were ob-
served on several natural samples. Figure 9. Roughness indices Z2, Dvar and Dyard for lab generated
fractures. The average value for each category is noted by a col-
oured dot, vertical lines show standard variation to the average.

667
Engineering in Chalk
Figure 10. Roughness indices Z2, Dvar and Dyard for natural frac-
ture planes (all samples but P10 fault). The average value for each
category is noted by a coloured dot, vertical lines show standard
variation to the average.
6 GENERAL CONCLUSIONS
An extensive process was established to collect, vis-
ually examine, classify, and scan natural and lab gen-
erated fractures. Observations within the geological
context were complemented by detailed surface char-
acterization combining (1) the empirical identifica-
tion of visual features and their depiction in terms of
JRC, with (2) analytical characterization by use of
statistical parameters and fractal dimensions. A clas-
sification method is introduced to allow for compari-
son between small sample surfaces. While the JRC
does not provide the best picture of small chalk fea-
tures due to scale effects, variability of texture and
surface geometry, Z2, Dyard and Dvar seem to provide
reasonable description of the small scale features ex-
amined (fault striation, twist hackles, ribs and plu-
mose).
The lab generated surfaces present similarities
with natural features: striation on shear planes, ribs,
and plumose-like features on tensile planes. They
both show comparable roughness (same order of
magnitude), with higher roughness on lab generated
fractures.
This preliminary study shows that the topic is
complex and further research is needed to deeper un-
derstand the characterization of the already sampled
geological formations with a more robust approach.
A more detailed analysis of the roughness indices
should be conducted. The surface classification
method for the 30x30mm samples may be refined.
668
Additional scans could be conducted on stratigraphic
joints and additional fault fractures. No direct link
between roughness and tectonic context stands out at
this stage. This question could be further examined
to look for similarities between the samples within
given tectonics backgrounds.
ACKNOWLEDGEMENTS
The authors would like to thanks Damien Bury for
the outstanding tech support, and all staff who helped
collecting rock samples on sites. SV is Research As-
sociate of the FNRS Belgium.
REFERENCES
Barton, N. & Choubey, V. 1977. The shear strength of rock joints
in theory and practice, Rock Mechanics and Rock Engineering 10,
1-54.
Brown S.R., 1987. Fluid flow through rock joints: effects of sur-
face roughness. J. Geophys. Research. 92 (B2) 1337-1347.
Develi, K. & Babadagli, T, Quantification of Natural Fracture Sur-
faces Using Fractal Geometry 1998. Mathematical Geology 30(8),
971-998.
Fairhurst, C & Hudson, J. 1999. Draft ISRM suggested method for
the complete stress-strain curve for intact rock in uniaxial com-
pression. International Journal of Rock Mechanics and Mining
Sciences 36, 279-289.
Faÿ-Gomord, O. 2017. Tight Chalks. Sedimentary and diagenetic
control on the petrophysical and fracturing properties. Phd thesis.
9.
Ge, Y. Kulatilake, P. Tang, & H. Xiong, C. 2013. Investigation of
natural rock joint roughness, Computers and Geotechnics 55, 290–
305.
ISRM, 1978. Suggested methods for Determining Tensile Strength
of Rock Materials. International Journal Mechanical Mining Sci-
ences & Geomechanics 15, 99-103.
Mandelbrot, BB. 1967. How long is the coast of Britain? Statistical
self-similarity and fractional dimension. Science 156, 636–8.
Nemoto, K. Watanabe, N. & Hirano, N. Direct measurement of
contact area and stress dependence of anisotropic flow through
rock fracture with heterogeneous aperture distribution. Earth and
Planetary Science Letters 281, 81–87
Nigon, B. Englert, A. Pascal, C. & Saintot, A. 2017. Multiscale
characterization of joint surface roughness. AGU Journal of Geo-
physical Research: Solid Earth 122, 1-15.
Vandycke, S. 2002. Paleostress records in Cretaceous formations
in NW Europe: extensional and strike-slip events in relationships
with Cretaceous-Tertiary inversion tectonics. Tectonophysics 357,
119 – 136.
Vogler, D. Walsh, S. & Bayer, P. 2017. Comparison of Surface
Properties in Natural and Artificially Generated Fractures in a
Crystalline Rock. Rock Mechanics and Rock Engineering 50, 11,
2891–2909.