Download pdf - Cilona Et Al JSG 2014

8/10/2019 Cilona Et Al JSG 2014

1/19

The effects of rock heterogeneity on compaction localization in porous

carbonates

Antonino Cilona a , *, Daniel Roy Faulkner b , Emanuele Tondi c, Fabrizio Agosta d,Lucia Mancini e, Andrea Rustichelli c, Patrick Baud f, Sergio Vinciguerra g

a Department of Geological and Environmental Sciences, Stanford University, Stanford, CA 94305, United Statesb University of Liverpool, Liverpool, United Kingdomc School of Environmental Sciences, University of Camerino, Camerino, Italyd Department of Science, University of Basilicata, Potenza, Italye Elettrae Sincrotrone Trieste SCpA, SS 14, Km 163,5 in Area Science Park, 34012 Basovizza (Trieste), Italy

f EOST Strasbourg, Strasbourg, Franceg Department of Geology, University of Leicester, Leicester, United Kingdom

a r t i c l e i n f o

Article history:

Received 18 December 2013

Received in revised form

8 July 2014

Accepted 16 July 2014

Available online 25 July 2014

Keywords:

Discrete compaction bands

Grain sorting

Triaxial compaction experimentsMechanical twinning

Porosity reduction

a b s t r a c t

Recent eld-based studies document the presence of bed-parallel compaction bands within the

Oligocene-Miocene carbonates of Bolognano Formation exposed at the Majella Mountain of central Italy.

These compaction bands are interpreted as burial-related structures, which accommodate volumetric

strain by means of grain rotation/sliding, grain crushing, intergranular pressure solution and pore

collapse.

In order to constrain the pressure conditions at which these compaction bands formed, and investigate

the role exerted by rock heterogeneity (grain and pore size and cement amount) on compaction local-

ization, we carried out a suite of triaxial compression experiments, under dry conditions and room

temperature on representative host rock samples of the Bolognano Formation. The experiments were

performed at conning pressures that are proxy of those experienced by the rock during burial (5e35 MPa). Cylinders were cored out from a sample of the carbonate lithofacies most commonly affected

by natural compaction bands. Natural structures were sampled and compared to the laboratory ones.

During the experiments, the samples displayed shear-enhanced compaction and strain hardening

associated with various patterns of strain localization. The brittleeductile transition occurred at 12.5 MPa

whereas compaction bands nucleated at 25 MPa conning pressure. A positive correlation between

conning pressure and the angle formed by the deformation bands and the major principal stress axis

was documented. Additional experiments were performed at 25 MPa on specimens cored oblique

(parallel and at 45) to the bedding. Detailed microstructural analyses, performed on pristine and

deformed rocks by using optical microscopy, scanning electron microscopy and X-ray computed

microtomography techniques, showed that grain crushing and mechanical twinning are the dominant

deformation processes in the laboratory structures. Conversely, pressure solution appears to be dominant

in the natural compaction bands. Experimental results highlight the strong inuence exerted by

bedding-parallel rock heterogeneity on both orientation and kinematics of deformation bands in the

studied carbonates.

2014 Elsevier Ltd. All rights reserved.

1. Introduction

Porous rocks form important reservoirs for water, hydrocarbons

and, potentially, the storage of greenhouse gases. Post-depositional

processes (i.e., mechanical, chemical, physical and biological) may

strongly affect their uid ow properties and, hence, are important

to determine. For this reason, the analyses of both deformation

mechanisms and ow properties of siliciclastic porous rocks have

received a good deal of attention, both in the eld (e.g.,Aydin et al.,

2006; Fossen et al., 2007) and in the laboratory (e.g., Wong et al.,

1997; Wong and Baud, 2012). Less attention has been paid to

porous carbonate rocks, which still constitute a large proportion of

oil reservoirs.* Corresponding author. Tel.: 1 6507259355, 1 6502007418 (mobile).

E-mail address:[email protected](A. Cilona).

Contents lists available atScienceDirect

Journal of Structural Geology

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c om / l o c a t e / j s g

http://dx.doi.org/10.1016/j.jsg.2014.07.008

0191-8141/

2014 Elsevier Ltd. All rights reserved.

Journal of Structural Geology 67 (2014) 75e93
mailto:[email protected]://www.sciencedirect.com/science/journal/01918141http://www.elsevier.com/locate/jsghttp://dx.doi.org/10.1016/j.jsg.2014.07.008http://dx.doi.org/10.1016/j.jsg.2014.07.008http://dx.doi.org/10.1016/j.jsg.2014.07.008http://dx.doi.org/10.1016/j.jsg.2014.07.008http://dx.doi.org/10.1016/j.jsg.2014.07.008http://dx.doi.org/10.1016/j.jsg.2014.07.008http://www.elsevier.com/locate/jsghttp://www.sciencedirect.com/science/journal/01918141http://crossmark.crossref.org/dialog/?doi=10.1016/j.jsg.2014.07.008&domain=pdfmailto:[email protected]


2/19

In carbonates, as well as any other rock, strain localization on

both small-scale laboratory samples and large-scale crustal fault

zones, signicantly inuences the stress eld (Paterson and Wong,

2005), strain partitioning (Olsson, 1999; Issen and Rudnicki, 2000)

and uid-transport properties of the deformed rocks (see Aydin,

2000; Faulkner et al., 2010a for full reviews). Previous studies

accurately investigated the deformation mechanisms associated to

the formation faults and fractures in carbonates by means ofeld

and/or laboratory analyses (e.g., Baud et al., 2000; Agosta et al.,

2007; Antonellini et al., 2008). Since these strain localizations are

mainly associated with dilatancy, scientic attention has focused

on the study of this phenomenon in tight carbonates. In contrast,

systematic eld and laboratory investigations of compaction-

assisted strain localization have been carried out only in recent

times.

Field-based studies (Tondi et al., 2006; Agosta et al., 2009;

Cilona et al., 2012; Rustichelli et al., 2012; Tondi et al., 2012;

Antonellini et al., 2014) described compaction localization within

limestones characterized by a wide range of porosities

(15 < f < 45%). These authors documented strain localization

occurring along narrow tabular bands oriented oblique and parallel

to bedding (i.e. compactive shear bands and compaction bands,

respectively; sensu Aydin et al., 2006). Bed-parallel compactionbands are characterized by a local porosity reduction and do not

show any macroscopic shear offset (Aydin et al., 2006). Based upon

the results ofeld and microstructural analyses, these compaction

bands have been interpreted as burial-related structures which

accommodate volumetric strain by the complex interplay of grain

rotation/sliding, grain crushing, intergranular pressure solution

and pore collapse. With the exception of intergranular pressure

solution, these micromechanisms are analogous to those previously

described in sandstones (e.g., Mollema and Antonellini, 1996; Aydin

et al., 2006; Fossen et al., 2007; Aydin and Ahmadov, 2009;

Eichhubl et al., 2010; Shultz et al., 2010; Deng and Aydin, 2012;

Alikarami et al., 2013).

Laboratory-based studies have investigated the parameters

controlling the mechanics of compaction of porous carbonates(10< f < 46%). Among these parameters, attention has been drawn

on uids type and/or chemistry (e.g., Homand and Shao, 2000;

Risnes et al., 2005; Zhang and Spiers, 2005 Croizeet al., 2010b),

temperature (e.g.,Croizeet al., 2010a) and pore size/type (e.g.,Zhu

et al., 2010). Other studies described the evolution of failure modes,

microstructures and micromechanism at different conning pres-

sures (Vajdova et al., 2004; Baxevanis et al., 2006; Baud et al., 2009;

Dautriat et al., 2011b; Cilona et al., 2012; Vajdova et al., 2012). In

contrast to sandstones for which compaction localization has been

reproduced in laboratory (e.g., Besuelle, 2001; Cuss et al., 2003;

Vajdova and Wong, 2003; Baud et al., 2004; Fortin et al., 2005;

Louis et al., 2006, 2009), porous carbonates mainly showed ho-

mogeneous deformation associated with the ductile regime (e.g.,

Vajdova et al., 2004, 2010; Baud et al., 2009; Zhu et al., 2010;Dautriat et al., 2011b; Vajdova et al., 2012). A recent pilot study

described compaction bands formed during laboratory experi-

ments performed on a carbonate grainstone under wet conditions

(Cilona et al., 2012). These authors suggested that a more system-

atic work was needed to understand fully the parameters control-

ling compaction localization in these rocks.

In this paper we perform a systematic set of triaxial compaction

experiments on specimens cored from the Bolognano Fm.

(Crescenti et al. 1969) to constrain the conditions at which natural

compaction bands nucleate and to investigate the role of rock

heterogeneity on compaction localization in porous carbonates.

The study rocks crop out at the Majella Mountain (central Italy) and

have been lately investigated by Rustichelli et al. (2012)and(2013).

The authors correlated compositional, sedimentological and pore

network characteristics to development and distribution of bed-

parallel compaction bands. In this work, an attempt to replicate

in the laboratory these natural features is performed. Dry experi-

ments, conducted under a range of conning pressures (5e35 MPa)

on samples cored in different orientations with respect to the

bedding (i.e., perpendicular, parallel and at 45 degree), are aimed

at evaluating the role exerted by primary rock constituents (e.g.,

grain and pore size/shape and cement amount/type) on any

compaction localization. The effect of rock heterogeneity and

anisotropy on the strength and failure modes of sandstones has

been previously studied by Louis et al.(2009) and Baud et al. (2012).

Both intact and deformed samples are characterized by means of

detailed microstructural analyses, performed by integrating optical

microscopy, Scanning Electron Microscopy (SEM) and X-ray

computed microtomography (mCT) techniques (Baker et al., 2012).

The results of microstructural analysis of experimentally-deformed

rocks arethen compared withthoseobtained from naturalstructures.

2. Methodology

2.1. Laboratory experiments

The experiments were carried out at the Rock DeformationLaboratory of Liverpool University. Nine cylindrical specimens

(20 mm diameter and 50 mm length) were cored, perpendicular to

bedding, from a boulder-sized hand sample. Moreover, two addi-

tional specimens (same shape) were cored parallel and oblique

(45) with respect to bedding. Each specimen was ground, to make

its bases parallel to each other, and then oven-dried for 48 h at

80 C to eliminate any residual humidity. The starting porosity

values were determined by means of helium multipycnometer

measurements using Quantachrome Instruments (MVP D150E).

The pore-throats distribution of one intact sample was computed

by mercury injection. The specimens were then placed in a 3.4 mm-

thick PVC jacket, in order to isolate them from the conning me-

dium (silicon oil). Eleven specimens were deformed in a conven-

tional triaxial conuration, under dry conditions and at roomtemperature, at conning pressures ranging from 5 to 35 MPa (cf.

Faulkner et al., 2010bfor details on the experimental apparatus).

The axial force applied on the sample was measured by an in-

ternal force gauge with a 0.02 kN resolution. The axial displace-

ment was measured using a linear variable differential transformer

(LVDT) attached, outside of the pressure vessel, to the electro-

mechanical servo-controlled ram for axial loading. The axial load

was applied at a xed rate of 0.5 mm s1 that corresponds to a

nominal strain rate of 105 s1. Since the studied rock was too

porous to enable the use of strain gauges, the volumetric strain of

the samples was recorded with a conning pressure volumometer

with a 0.1 mm3 resolution. The volumometer was calibrated by

loading a steel blank specimen up-to 20 kN while keeping the

conning pressure constant.

2.2. X-ray microtomography (mCT)

A selection of samples (i.e. deformed and pristine) was vacuum

impregnated with epoxy resin and imaged at the Elettra synchro-

tron light laboratory in Basovizza (Trieste, Italy) by two different

instruments. Each cylindrical sample (diameter of 20 mm) was cut

in two parts: one half-cylinder was imaged by conventional

microfocus X-raymCT at the TomoLab station (Zandomeneghi et al.,

2010). A smaller parallelepiped-shaped sample was cut from the

remaining half cylinder to be investigated by using phase-contrast

synchrotron radiation (SR) mCT at the SYRMEP beamline (Tromba

et al., 2010). Details about samples investigated by X-ray mCT are

reported inTable 1.

A. Cilona et al. / Journal of Structural Geology 67 (2014) 75e9376


3/19

At SYRMEP, the SR is provided by a bending magnet source; the

sample is placed at a distance of about 23 m from the source and is

illuminated by a monochromatic, nearly parallel X-ray beam with a

maximum area of about 160 6 mm. The monochromator is based

on a double Si(111) crystal system making it possible to tune the X-

ray energy in the range 8.3e38 keV. The high spatial coherence of

the X-ray beam permits to benet of phase-contrast effects

employing a free-space propagation based technique (Cloetens

et al.., 1996). The sample projections were recorded with a 12 bit,

water-cooled CCD camera (4008 x 2672 pixels, 4.5 4.5 mm

effective pixel size). The SR mCT scans were acquired in the

following conditions: X-ray energy 34 keV, sample-to-detector

distance 180 mm, exposure time/projection 4.8 s, pixel

size9.0 mm. For each scan 1440 projections were recorded over a

180 rotation. The reconstruction of the slices was performed using

the SYRMEP_Tomo_Project 4.0 software and the GRIDREC algo-

rithm (Dowd et al., 1999) with an isotropic voxel size of 9.0 mm.

The TomoLab instrument is equipped with a sealed microfocus

source (Voltage range: 40e130 kV, maximum current 300 mA,

minimum focal spot 5 mm) and it is based on a cone-beam ge-

ometry. A water-cooled, 12 bit CCD camera was employed as de-

tector (4008 2672 pixels, effective pixel size 12.5 12.5 mm2). The

sample was imaged in the following conditions: Voltage 130 kV,current 61 mA, 1.5 mm thick Al lter, exposure time/

projection 5.8 s, pixel size 17.2 mm. For each scan 2400 pro-

jections were recorded over a 360 angular rotation. The com-

mercial software COBRA 7.4 (Exxim) was used for slice

reconstruction with an isotropic voxel size of 17.2 mm. By using this

software, a beam-hardening correction based on a polynomial

tting of the measured nonlinear relationship between object

thickness and the log ratio of the intensities was used for slice

reconstruction. In this study, the use of two different mCT in-

struments allows a multi-scale and complementary analysis of the

imaged samples. Phase-contrast SR mCT technique outperforms

conventional mCT in pore space determination. This is due to the

higher spatial and contrast resolution related to the use of a nearly-

parallel, monochromatic high-intensity X-ray beam leading toreconstructed slices free of beam hardening and magnication ef-

fects (Kak and Slaney, 1988) and with a high signal-to-noise ratio.

Moreover, by working in phase-contrast mode (edge-detection

regime), all the interfaces corresponding to phase changes in the

sample show an enhanced visibility (Cloetens et al., 1996; Mancini,

1998). Compared with absorption images, the objects appear

sharper and very small details can be detected (also smaller than

the pixel size of the detector) Cloetens et al. (1997). The epoxy,

partially invading the pores of the sample, could be visualized and

easily distinguished from the empty space (within the limit of

detail detectability of the employed technique) giving a higher

accuracy in the image analysis procedures used to separate the

solid phase from the porous space. However, conventional mCT

imaging allowed us to obtain a quantitative characterization ofmacroporosity (pores with sizes of the order or larger than 30 mm)

on a larger volume than SRmCT (Table 1). The quantitative analysis

of the volumes was carried out by means of the Pore3D software

library developed at Elettra (Brun et al., 2010; Zandomeneghi et al.,

2010). The same software was used to lter the slices reconstructed

by conventionalmCT in order to reduce the ring artefacts present in

the images.

The reconstructed slices were visualized by using the ImageJ

software (Abramoff et al., 2004) while the volume renderings of

raw and processed images were obtained by the commercial soft-

ware VGStudio MAX 2.0.

In order to separate the different phases present in the sample, a

segmentation process has to be applied to the reconstructed vol-

umes. In the literature, several authors treated the problem to es-

timate macro- and micro-porosity from X-ray mCT images in

carbonates spanning across several tens of length scales (Bauer

et al., 2011;Blunt et al., 2013; Wildenschild et al., 2013). Different

approaches have been proposed based on multi-scale analysis

combining mCT with SEM imaging and/or on ltering and seg-

mentation of mCT images into three phases. The three resulting

phases represent the resolvable porous space (macro-porosity), the

micro-porosity (below the resolution of the mCT image), and the

solid region, respectively (Sok et al., 2009; Ji et al., 2012).

We distributed the voxels in two phases, i.e. solid and void, by

manually selecting a threshold value in 3D from the histogram of

the intensities (Zandomeneghi et al., 2010) in the analysed Volume

Of Interest (VOI). In this way, gray scale images were converted into

a binary images from which it was possible to estimate the 3Dporosity of the samples. Because this approach allows to quanti-

tative characterize only the pores with sizes compatible with the

spatial resolution of the applied mCT instrument, we integrated this

technique with image analysis of SEM images.

2.3. Microstructural analysis

The epoxy-impregnated samples were cut along a plane parallel

to the axial direction to prepare petrographic thin sections. The thin

sections were polished to be analysed using an optical polarizing

microscope (Nikon Eclipse E600 Pol microscope) and eventually

carbon coated to be observed under Scanning Electron Microscope

(JEOL JXA-8230). 2D porosity values were obtained on micropho-

tographs (6 mm pixel size) by means of the open source softwareImageJ 1.32, (see Rustichelli et al., 2012 for details on the meth-

odology). Quantitative microstructural analysis was carried out on

representative samples to determine micro-cracks density. After

identifying the portions of the samples where localization

occurred, we selected 88 microphotographs (3 mm of pixel size)

representing different degree of deformation. A rectangular grid of

9 mm2 (3.25 2.75) of area was superimposed on each analyzed

microphotograph, using stereological techniques, the number of

micro-crack intersections with a test array of 15 parallel lines

spaced at 0.2 mm was manually counted. Measurements were

made in two orthogonal directions parallel and perpendicular to

the s1 axis, respectively. We denoted the linear intercept density

(number of micro-crack intersections per unit length) for the array

oriented parallel to s1 by PkL , and that for the perpendiculararray by

PL . Previous studies (Underwood, 1970) have demonstrated that

since the spatial distribution of damage is approximately axisym-

metric in a triaxially deformed sample, the crack surface area per

unit volume (SV) can be inferred from linear intercept measure-

ments along two orthogonal directions, see Equation(1):

Table 1

List of the analyzed samples and the acquisition parameters.

Sample code Description Facility Voxel size [mm] Imaged volume: geometry Imaged volume [mm] VOI [mm3]

BoloH Host rock TomoLab 17.2 Half cylinder Diameter 20; Height 21 6.7 5.1 3.0

Natural CB Compaction Band SYRMEP 9 Parallelepiped 4 4 8 3.5 3.8 3.7

Bolo5 Deformed at 25 MPa Pc TomoLab 17.2 Half cylinder Diameter 20; Height 21 4.4 10.1 4.3

Bolo5 Deformed at 25 MPa Pc SYRMEP 9 Parallelepiped 4.5 4.5 8.0 3.8 3.8 2.3

A. Cilona et al. / Journal of Structural Geology 67 (2014) 75e93 77


4/19

Fig.1. Geological setting of Majella Mountain. a) Schematic geological map (modied afterGhisetti and Vezzani, 1998). b) Stratigraphic scheme of the carbonate succession (modie

(modied afterVezzani and Ghisetti, 1998).


5/19

Sv P

2PL

2

P

2

P

jL (1)

In addition, we used the Equation(2) to dene the value of the

micro-crack anisotropy factor (U23,Underwood, 1970):

U23 PL P

jL

PL 4=P 1PjL

(2)

The parameterU23 represents the ratio between the surface area

of micro-cracks parallel to s1and the total micro-crack area.

The density of twinned calcite grains was obtained using the

methodology described byVajdova et al. (2010). Microphotographs

of intact and deformed samples (i.e., localization zones and sur-

rounding parts) were selected, an array of squares (#35) was then

superimposed on each picture (#39 in total, 6 mm pixel size). The

number of twinned calcite crystal was counted within each square

(area of 1 mm2) and mean values were obtained for the localization

zones and the surrounding parts of the specimens.

3. Geological framework

The Majella Mountain is an east-verging, thrust-related anti-

cline located in the external zone of the central Apennines, central

Italy (Fig. 1;Ghisetti and Vezzani, 2002). The stratigraphic succes-

sion includes a 2 km-thick sequence of Cretaceous to Miocene

Fig. 2. a) Backscattered Electron image of the skeletal grainstones of the facies B, legend: Bryozoan fragment (B), Echinoid plate (E), Syntaxial overgrowth cement (S), intergranular

(Pi) and intragranular (Pii) porosities. Laminae dominated by intergranular or intragranular pores are highlighted by red and blue box, respectively. b) Pore-throat statistics from

mercury injection on a host rock sample; Microfocus X-raymCT images (voxel size 17.2mm) of intact Bolognano grainstones. c) an example of a 2D axial slice; d) volume rendering

of a parallelepiped-shaped volume (height in the zdirection 21 mm) showing both pores (black) and grains plus cement (grey); e) volume rendering of the same sample region

illustrated in d) after segmentation: only the pores are shown. The porosity values computed within the blue and red polygons are also reported. (For interpretation of the ref-

erences to colour in this gure legend, the reader is referred to the web version of this article.)



6/19

carbonate formations related to various depositional settings

(platform and slope/ramp) originally pertaining to the northern-

most sector of the Apulian Platform realm (Vecsei et al., 1998). The

Apulian carbonates are overlain by marine siliciclastic and evapo-

ritic deposits of Messinian-to-Pleistocene age, the majority of

which originally deposited within the Peri-Adriatic foredeep basin

of the central Apennines fold-and-thrust belt (Vezzani and Ghisetti,

1998). According toGhisetti and Vezzani (2002)and Agosta et al.

(2009), the development of the Majella thrust-related anticline

occurred during the Middle-to-Late Pliocene. The internal defor-

mation of this box-shaped fold, characterized by two steeply-

dipping to overturned limbs, mainly consists of high-angle

normal, strike-slip and oblique-slip faults, small folds and

different sets of mode I fractures (e.g., Marchegiani et al., 2006;

Tondi et al., 2006; Antonellini et al., 2008; Agosta et al., 2009,

2010; Aydin et al., 2010). Conversely, based upon crosscutting re-

lationships with Messinian-to-Pleistocene sediments topping the

carbonates, Scisciani et al. (2002) interpreted faulting activity as

Messinian.

We analyzed structures pertaining to the Oligocene-Miocene

ramp carbonates of the Bolognano Formation cropping out in the

Roman Valley Quarry area of the MajellaMt. (Fig.1). The carbonates

of the Bolognano Formation consist of i) relatively shallow-waterskeletal grainstones and packstones, which are made up of frag-

ments of Larger Benthic Foraminifera (LBF), bryozoans, red algae

and lamellibranches, echinoid plates and spines, and ii) deeper

water marly wackestones and mudstones with planktonic forami-

nifera (Fig. 2a; Vecsei and Sanders, 1999; Pomar et al., 2004;

Rustichelli et al., 2013). These marine carbonates accumulated on

an isolated, gently dipping, carbonate ramp under sub-tropical

conditions and nutrient-rich sea water (Westphal et al., 2010;

Rustichelli et al., 2013).

For the laboratory experiments, we selected the lithofacies B in

Rustichelli et al.'s (2013)stratigraphic section because it is the one

that is most densely crosscut by natural bed-parallel compaction

bands (Rustichelli et al., 2012). This lithofacies consists of medium-

grained (mean grain size0.3 mm), moderatee

(Fig. 2a blue box)to well-sorted (Fig. 2a red box) skeletal grainstones composed by

more than 99.5% of calcite (Rustichelli et al., 2012). Some of these

grainstones contain hydrocarbon residues in the form of bitumen

(Agosta et al., 2009). This lithofacies is dominated by bryozoan

fragments and minor amounts of LBF (mainly Amphistegina and

Operculina, rareMyogipsina), echinoid plates and spines, red algae

and lamellibranch fragments (Fig. 2a;Rustichelli et al., 2012; 2013).

The lithofacies is upto 15 m-thick, and it is madeup of stacks of 10 s

of cm- to 4 m-thick crossbed packages bounded by sub-horizontal

truncation surfaces (Rustichelli et al., 2012; 2013).

The pore-throats distribution, obtained from mercury injection

performed on one intact sample (26 mm in diameter and 29 mm in

height) and interpreted based on thin section analyses, revealed a

bi-modal porosity: larger (0.1e

0.01 mm) intergranular pores andsmaller (0.01e0.001 mm) intragranular pores (Fig. 2b).Rustichelli

et al. (2012) showed that most of the intragranular micro-pores

(10 s-to-1 mm) are localized within echinoid and bryozoan frag-

ments, additional intergranular micro-pores are encompassed be-

tween microscparry cement particles. To quantify the amount of

intragranular and intergranular micro-pores, we applied image

analysis techniques on high-magnication (150) SEM images

(0.5 mm pixel size). Within the echinoid and bryozoans fragments

we measured an average porosity of 5.5% 0.95 (standard devia-

tion), whereas we measured an average porosity value of

51.5% 5.4 (standard deviation) between the microsparry cement

particles. Because echinoid and bryozoans fragments represent on

average the 60% of the grains in the studied lithofacies and the

microsparry cement isz

1.2% of the of the rock volume (Rustichelli

et al., 2012), we estimated that z4% of the total rock volume is

constituted by micro-pores not detectable by the SR mCT facility

(Table 1).

A host rock sample was imaged by microfocus X-ray mCT and

porosity values were obtained from selected VOIs (seeTable 1). At a

sample scale the host rock shows rhythmic alternation of laminae

(0.2e2 cm thick) where larger pores are dominant (red polygon,

Fig. 2e) and layers where smaller pores are more abundant (blue

polygon Fig. 2e). The presence of different pore-size distributions is

responsible of the different porosity values (up to 4% of difference)

measured within different laminae sub-parallel to bedding

(Fig. 2dee).

4. Natural bed-parallel compaction bands

At the Roman Valley Quarry the bed-parallel compaction bands

(CBs) dip 10-15 towards north and represent the rst structures

that developed within the studied carbonates (Agosta et al., 2009;

Rustichelli et al., 2012). At outcrop scale, the bed-parallel

compaction bands appear lighter-coloured with respect to the

surrounding host rock. In thin section, the inner texture of the CBs

consists of tight fossil fragments and a very little amountof residual

porosity mainly localized within the single grains (Fig. 3f). Acompaction band was investigated by phase-contrast SR mCT

(Table 1), an axial slice and the volume rendering of a sub-volume

are shown inFig. 3cee. A porosity ofca. 12% was calculated within

the CB from a segmented VOI, whereas adjacent to the CB the

measured porosity was ca. 16%. Across the thickest compaction

bands, a progressive grain-size reduction was observed from the

external to the more internal portion of the structure, in agreement

withCilona et al. (2012). The grain-size reduction was determined

by the interplay of intergranular pressure solution and Hertzian

cracking (Zhang et al., 1990) that occurred at the grain-to-grain

contacts (Fig. 3f). Adjacent and within individual compaction

bands, calcite twinning was observed on overgrowth syntaxial

cement, which behave like a single crystal.

5. Laboratory deformation experiments

Experiments were conducted on samples cored perpendicular

to bedding at various conning pressures (5e35 MPa) with the aim

of determining the yield points for the Bolognano Formation.

Microstructural investigations were performed to identify if any

compaction localization had occurred, or whether the deformation

was distributed. As will be shown later, compaction localization did

occur within certain pressure conditions. Then, based on the rst

results, we performed further experiments at the pressure corre-

sponding to compaction localization, using samples cored at

different orientation with respect to the bedding. These latter ex-

periments were useful to discuss possible the effect of rock het-

erogeneity on the compaction localization.

5.1. Bedding-perpendicular samples

5.1.1. Mechanical data

In the following section, we consider the compressive stresses

and compactive strains (i.e., shortening and porosity decrease) as

positive. The maximum and minimum (compressive) principal

stress axes are denoted by s1and s3, respectively. The mean stress

(s1 2s3)/3 and the differential stresss1 e s3 are denoted by Pand

Q, respectively. We dene as brittle a failure mode in which sig-

nicant strain softening is recorded immediately following the

peak stress. We classify as ductile a failure mode in which a

stressestrain curve displays strain hardening immediately

following yield (cf.Jaeger et al. 2006).



7/19

Mechanical data pertaining to the experiments (Table 2), are

shown inFig. 4. The differential stress as a function of axial strain is

shown in two different graphs: in Fig. 4(a) are the results of ex-

periments carried out at a conning pressure (Pc) comprised be-

tween 5 and 12.5 MPa; in Fig.4(c) the results of experiments

conducted at Pc ranging between 15 and 35 MPa. Volumetric strain

as function of the mean stress is also displayed in two other graphs:

inFig. 4(b) are shown the results of experiments with Pc ranging

between 5 and 12.5 MPa; inFig. 4(d) those from experiments with

Pc comprised between 15 and 35 MPa.

Because it was not possible to measure volumetric strain with

strain gauges, the onset of pore collapse under hydrostatic

compression P* (see Wong et al., 1997) could not be directly

measured on dry Bolognano grainstone. To circumvent this prob-

lem, we followed the procedure applied byBaud et al. (2009) and

performed another test at a constant differential stress of 3 MPa.

The conning pressure was slowly increased and the onset of pore

collapse (55 MPa) could be discerned on the axial strain measured

by a DCDT (Direct Current Differential Transformer; Fig 4e). This

test allowed us to infer P* to be 58 MPa (Fig. 4f).

The experiments performed at 5e12.5 MPa Pc, showed typical

brittle behaviour (Fig. 4a); the stressestrain curves reached a peak

beyond which strain softening followed. Two experiments were

performed at 5 MPa Pc and these were in excellent agreement

Fig. 3. Compaction bands (CB) parallel to bedding (B) analyzed along the walls of the Roman Valley Quarry. aeb) CB affecting the grainstones of the lithofaciesB. Pen is for scale. The

grainstones are strongly invaded by hydrocarbons which highlights the whitish CB (afterRustichelli et al., 2012). Phase-contrast sy nchrotron X-raymCT images (voxel size 9.0mm):

c) 2D axial slice showing grains, cement, pores and the resin inside the pores; d) volume rendering of a sub-volume (size 2.7 3.6 4.3) mm3 showing both pores (black) and grains

plus cement (grey); e) volume rendering of the corresponding segmented volume where the pores are illustrated, the dashed contour highlights the CB. f) Backscattered Electron

images collage of a CB (red dashed), pressure solution and grain crushing are responsible of the local grain-size reduction. (For interpretation of the references to colour in this gurelegend, the reader is referred to the web version of this article.)



8/19

(Fig. 4a). While the peak stresses were quasi-constant (53 MPa) for

both experiments, some variability in the post-peak stress drops

was observed (Fig. 4a).

The experiment performed at 12.5 MPa Pc did not show a sig-

nicant stress drop after the peak stress, just a gentle strain soft-

ening (Fig. 4a). This behaviour suggests that around this Pc the

brittle/ductile transition occurs (Paterson and Wong, 2005). Beyond

15 MPa conning pressure, the specimens experienced ductile

failure (Fig. 4c). Although the results of the latter experiments

documented various stress vs. strain curves shapes, in all cases

stress drops of a few MPa were recorded in the post-peak phase,

which is generally consistent with the occurrence of strain locali-

zation (Vajdova and Wong, 2003; Baud et al., 2004).

The volumetric strain data indicate that the mechanical

behaviour of the deformed carbonate specimens was always

compactant (Fig. 4b and d). This result is in agreement with those

previously obtained after dry experiments on carbonates with

comparable porosity values (Baud et al., 2009; Zhu et al., 2010;

Dautriat et al., 2011b).

Wong et al. (1997) showed that the hydrostatic and non-

hydrostatic loadings are coupled in a triaxial compression

experiment. In the compactive regime, these authors identiedthe yield point as the critical pressure C*, at which the deviatoric

part of the loading results in an acceleration of the compaction

with respect to the hydrostatic behaviour. Since we could not

directly compare the tests with the hydrostatic one, we deter-

mined theC* based on the change in slope of the stress vs. strain

curves (see Fig 4b as example). A shear-enhanced compaction

(sensu Curran and Carroll, 1979) behaviour was observed during

all of the performed experiments (Fig. 4bed). The computed

yield point values are reported in Table 2. InFig. 4(f) all experi-

mental data are shown in the stress space (mean stress vs. dif-

ferential stress), in order to document the yield points computed

for this carbonate lithofacies.

5.1.2. Structural analysis of experimentally-deformed specimens

5.1.2.1. Macroscopic observations. After being deformed, the sam-

ples were unloaded and then retrieved from the pressure vessel to

identify if strain localization occurred along deformation bands

(DBs). Deformation Bands appeared lighter-coloured with respect

to other portions of the specimens. The angles formed by the DBs

ands1varied according to the magnitude of the conning pressure

(Fig. 5). At 5 MPa Pc DBs formed angles of 40 and 30, respectively

(Fig. 5, Bolo1 and Bolo7). At 12.5 MPa of Pc, strain localization

occurred along a conjugate pair of DBs oriented at 53 and 77 with

respect to s1, respectively. At higher conning pressures (15 and

20 MPa), we documented the formation of clusters of sub-parallel

DBs, in agreement with Mair et al. (2000). At 15 MPa Pc, these

features were oriented at 55

with respect to s1, whereas they

formed an angle of 72 at 20 MPa Pc Sub-perpendicular (87) to s1DBs nucleated at 25 MPa Pc At the highest Pc used (35 MPa) strain

localization did not occur and we only documented homogeneous

deformation (Fig. 5). These results are in agreement with those

from previous studies conducted on carbonates (Baud et al. 2009;

Cilona et al., 2012) and sandstones (Baud et al., 2004) for which a

general positive correlation was found between the conning

pressure and the angle formed by DBs and the maximum

compression axis.

5.1.2.2. Microstructural analysis. During deformation, shear-

enhanced compaction behaviour was followed by the three main

failure modes: shear failure, compaction localization and homo-

geneous deformation. We present the results of ve selected

samples (i.e. Bolo H; Bolo7; Bolo11; Bolo5; Bolo9) that represent

different stages of deformation, from 0 to 3% of axial strain

(Table 2). With the exception of Bolo H (host rock), each analyzed

sample was subjected to a different failure mode: low-angle to s1compactive shear bands (Bolo7 and Bolo11); compaction bands

(Bolo5) and homogeneous deformation (Bolo9).

The analyzed DBs were less than 3 mm-thick, their internalstructure was characterized by heterogeneously cracked grains,

twinned calcite crystals and collapsed pores (Figs. 6 and 7).

Deformation bands presented a local porosity reduction (Fig. 8c) in

their internal portions associated with pore and grain size reduc-

tion (Figs. 6 and 7). Despite deformation bands generally appearing

planar and continuous features, up to 30% of thickness variation

was documented along the same DB (Fig. 6i). During the experi-

ments conducted in ductile regime the DBs generally tended to

cluster within laminae characterized by a better grain sorting,

larger pores and relatively richer in bryozoan fossils ( Figs. 6 and 7).

This result is in agreement with previous eld (Rustichelli et al.,

2012) and laboratory observations (Cheung et al., 2012).

Both distribution and density of cracks were calculated in

different portions of the samples (see Section2.3). The values ofdamage obtained for all the analyzed samples are relatively high

(e.g., Wu et al., 2000). Indeed the host rock, which we used as a

baseline value of the pre-experimental deformation damage,

showed a crack density of 6.8 and 8.3 parallel and perpendicular to

bedding, respectively. The anisotropy factor U23 in this sample is

0.14 meaning that a higher number of cracks is oriented perpen-

dicular to bedding.

All the samples, except for Bolo9, showed crack density values

two-to-three times higher than the values of the host rock ( Fig. 8).

Moreover, the crack density was higher within the regions where

strain localization occurred with respect to the surrounding parts of

the samples.

Bolo7 and Bolo5 have a more anisotropic crack distribution

within the strain localization areas than the surrounding parts.

Table 2

Summary of the triaxial experiments performed of the Bolognano grainstones.

Sample Angle to

bedding []

Porosity [%] Conning

pressure [MPa]

Axial strain [%] C* Microstructure

Diff. Stress [MPa] Mean stress [MPa] Respect tos1

Bolo1 90 27 5 0.8 44.7 19.9 Low angle

Bolo7 90 27 5 1.8 44.42 19.8 Low angle

Bolo11 90 26.3 12.5 3 49.150 28.88 Low angle high angle

Bolo3 90 28.3 15 3 48.175 31.06 High angle

Bolo4 90 28 20 2.85 42.24 34.08 Very high angle

Bolo5 90 27.3 25 2.85 40 38.3 Perpendicular

Bolo 45 to bed 45 26.7 25 2.85 49.07 41.35 45

Bolo 0 to bed 0 26.6 25 2.6 46.17 40.35 Low angle high angle

Bolo9 90 26.8 35 2 28.1 44.37 Homogeneous deformation

Bolo12 90 32 8 3 55 Homogeneous deformation



9/19

Conversely, Bolo11 displays a more isotropic crack distribution in

the localization area respect to the surrounding parts. Bolo9 shows

a marked anisotropy of the crack distribution, but the standard

deviation of these values is high.

The internal structure of sample Bolo5 was investigated by

means of X-raymCT both using the conventional and the SR sources

(seeTable 1for details). X-ray mCT data were used to calculate the

porosity within and outside the compaction band. The porosity

Fig. 4. Mechanical data for triaxial compression experiments on Bolognano grainstones. a) Differential stress versus axial strain for experiments at conning pressures up to

12.5 MPa. b) Mean stress versus volumetric strain for experiments at conning pressures up to 12.5 MPa. c) Differential stress versus axial strain for experiments at conning

pressures up to 35 MPa. d) Mean stress versus volumetric strain for experiments at conning pressures up to 35 MPa. e) Mean stress versus axial strain for a constant differential

stress experiment. f) Yield points of the experiments represented in the stress space, different symbols for different orientation respect to bedding.



10/19

values calculated within the compaction band are 16% with the

conventional mCT data (voxel size 17.2 mm) and 20% with SR mCT

data (voxel size 9 mm) (Fig. 9).

The porosity calculated outside the compaction band are 23%

and 25% with conventional and SRmCT, respectively. Relevant pore-

size reduction is documented within the laboratory compaction

band (Fig. 9bee). Within the localization zone, where small pores

are more abundant, the high accuracy of the phase-contrast SRmCT

technique (see Section2.2) gives a more precise approximation of

the porosity with respect to the conventional mCT technique (Fig. 9).

5.2. Bedding-oblique samples

With the aim of investigating possible effects of rock hetero-

geneity on strain localization, we cored two specimens at different

orientations with respect to sedimentary bedding (parallel and at

Fig. 5. Pictures of the deformed samples, in white the superimposed interpretation of the formed deformation bands.

Fig. 6. Compilation of photomicrographs (cross-polarized nicols) and SEM images from the samples Bolo H (a, b) Bolo7 (c, d, e, f) and Bolo11 (g, h, i). Red arrows indicate some of the

micro-cracks, T represents twinned calcite crystals, pores are black, DBs are evidenced with red dashed lines. In all the images of deformed samples the direction of s1 is

horizontal. a) micro-cracks; b) pore-emanated cracks. The images represent different portions of the samples: c and f are relatively-undeformed areas; d and h are areas adjacent to

deformation band; e and i show the deformation band. (For interpretation of the references to colour in this gure legend, the reader is referred to the web version of this article.)



11/19

Fig. 7. Compilation of photomicrographs (cross-polarized nicols) and SEM images from the samples Bolo5 (a, b, c) and Bolo9 (d, e, f). Red arrows indicate some of the micro-cracks,

T represents twinned calcite crystals, pores are black. In the images the direction of s1 is horizontal. The images represent different portions of the samples: a relatively-

undeformed area; b area adjacent to compaction band with twinned calcite; c) within the compaction band. d) twinning-rich zone, e) bryozoans-rich and twinning-free

portion, f) strength difference between echinoid (E) and bryozoans fragments. (For interpretation of the references to colour in this gure legend, the reader is referred to the

web version of this article.)

Fig. 8. Quantitative data ofSv(a); U23(b) and 2D porosity (c). Each color represents a different sample, the error bars represent the standard deviation. The data are divided in two

groups in order to differentiate the measures within the strain localization area from those in the rest of the sample. (For interpretation of the references to colour in this gure

legend, the reader is referred to the web version of this article.)



12/19

45). We run these two additional tests at 25 MPa Pc because it

corresponds to the pressure at whichs1-perpendicular compaction

bands formed (Fig. 5).

5.2.1. Mechanical data

The mechanical data are shown inFig. 10. The differential stress

as a function of axial strain is shown inFig.10(a), the mean stress as

the function of the volumetric strain inFig. 10(b). Those data arecompared to that obtained, at the same Pc, for the specimen cored

perpendicular to bedding.

For the bed-parallel specimen we documented a peak stress of

59 MPa followed by a 2 MPa-wide stress drop. On the specimen

cored at 45 to bedding, we recorded a peak stress of 61.5 MPa

before a stress drop of 8 MPa occurred. In this latter experiment

strain-hardening was more pronounced and further stress drops

were recorded in the post-peak part of the curve.

5.2.2. Structural analysis of bedding-oblique specimens

5.2.2.1. Macroscopic observations. As shown inFig. 11macroscopic

deformation was documented in both samples. On the specimen

cored at 45 to bedding, strain localization occurred and 45 to54-

to s1 DBs formed. Macroscopic shear offset (z0.3 mm) wasaccommodated along DBs parallel to the bedding (Fig. 11).

Discontinuous areas of deformation were documented on the bed-

parallel specimen as well. The general angle formed by these

patches of deformation and the s1 direction ranged between 65

and 85.

5.2.2.2. Microstructural analysis. During both experiments shear-

enhanced compaction was associated with two different failure

modes: strain localization and homogeneous damage. Both sam-

ples (i.e., Bolo 45 to bedding and Bolo 0 to bedding) were deformed

up to 3% axial strain (Table 2).

The thicknesses of the DBs documented within sample Bolo45

to bedding ranged from 1.7 to 3.3 mm (Fig. 11), their internal

structure showed highly-cracked grains, twinned calcite crystals

and collapsed pores (Fig. 12bec). Local porosity reduction, associ-

ated to pore and grain sizes reduction, was documented within the

DBs formed in Bolo45 to bedding (Fig. 8c). In this specimen the DBs

localized within the layers characterized by higher presence of

bryozoans fossils (Fig. 11). Based on the lower porosity of these DBs

with respect to the host rock (Fig. 11) and the macroscopic shear

offset they resolved, we classied them as compactive shear bands

(CSB;sensu Aydin et al., 2006).

The texture of sample Bolo0 to bedding was caused by diffusedeformation, strain localization did not occur in this sample but the

Fig. 9. aeb) Volume rendering of a parallelepiped-shaped volume (height of the volume in the zdirection 210 mm) of sample Bolo5 obtained by microfocus X-ray mCT (voxel

size 17.2mm): in a) both pores (black) and grains plus cement (grey) are visible. The red dashed polygon highlights the CB area and measured 3D porosity values are also shown. In

b) the corresponding segmented volume is visible with the pore network within and outside the compaction band. c) Volume rendering of a parallelepiped-shaped volume of

sample Bolo5 obtained by phase-contrast synchrotron X-raymCT (voxel size 9.0mm): the red dashed polygon highlights the CB area. Measured 3D porosity values are also shown.

e) Backscattered Electron images mosaic, red dashed polygons highlight discrete CBs. (For interpretation of the references to colour in this gure legend, the reader is referred to the

web version of this article.)



13/19

central part of the sample was highly deformed (Fig. 11). The

texture of this specimen was tighter than the other two samples

deformed at the same Pc condition (Fig. 11). Pore collapse and

mechanical twinning occurred within this sample. The mechanical

twinning of calcite crystals was enhanced where grains pile-up

occurred (Fig. 12d). For sample Bolo45 to bedding crack density

was up-to-two times higher than the host rock (Fig. 8). Highercrack

density was documented in the area of localization with respect to

the surrounding parts of the sample. Within the strain localization

area, the crack distribution was less isotropic than the surrounding

parts. The crack density values for sample Bolo0 to bedding were

slightly higher than the host rock (Fig. 8). Although the damagewas

homogeneously distributed and the crack distribution was aniso-

tropic, high standard deviation values of the crack distributionsuggest isotropy in some analyzed pictures and anisotropy in others

(Fig. 8).

Fig. 12(a) shows the results of twinning density by means of

column diagrams. Both samples, where strain localization took

place, presented higher twinning densities values within the

deformation bands with respect to surrounding portions of the

samples. Among the three samples the maximum twinning density

was record in the sample 45 to bedding. The minimum density of

twinning was found to correspond to the sample perpendicular to

bedding.

6. Discussion

6.1. Mechanical behaviour

The carbonate grainstones discussed in prior sections presented

post-peak stress drops in both brittle and ductile regimes (Fig. 4).

We documented a brittle behaviour up to 10 MPa Pc whereas the

brittle/ductile transition was observed at 12.5 MPa Baud et al.

(2009), in their experiments on carbonates with comparable

porosity, documented this transition at similar Pc conditions.

Ductile failure occurred beyond 12.5 MPa Pc associated to

strain-hardening. Beyond the peak stress, and during the strain-

hardening part of the curves, stress drops events were recorded:

the stress drops are generally associated to the occurrence of strain

localization (Baud et al., 2004). These observations are consistent

with the stressestrain curves of laboratory-deformed carbonates

described byCilona et al. (2012)and previous work on sandstones

(e.g., Vajdova and Wong, 2003; Baud et al., 2004; Fortin et al.,

2006).

Fig. 10.Mechanical data for triaxial compression experiments (performed at 25 MPa) on samples of Bolognano grainstones cored at different orientations with respect to bedding.

a) Differential stress versus axial strain. b) Mean stress versus volumetric strain.

Fig. 11. Pictures of the sample deformed at 25 MPa, in white the superimposed interpretation of the formed deformation bands. For the two oblique-to-bedding samples (Bolo45 to

bed and Bolo0 to bed) a mosaic of SEM images shows their internal texture. The red dashed lines highlight the CSB. (For interpretation of the references to colour in this gure

legend, the reader is referred to the web version of this article.)



14/19

Temperature may condition the mechanics of carbonates

(Paterson and Wong, 2005): higher temperatures generally pro-

mote ductility and increase the sensitivity of carbonates to strain

rate variations. Croize et al. (2010a) performed experiments on

bioclastic carbonate sand at temperatures up to 70 C (corre-

sponding to a depth >2 km, under a normal geothermal gradient:

25 C/km). Their results showed that within that range the tem-

perature does not affect the mechanical response of carbonate

sands under strain rates comparable to those of our experiments.

Based upon the latter evidence, we suggest that the mechanical

strength and failure behaviour of our samples, acquired at a room

temperature, were not affected by thermally activated deformation

mechanisms, in agreement withLiteanu et al. (2013).

Our data are coherent with a purely compactant mechanical

behaviour of Bolognano grainstones. Shear-enhanced compaction

behaviour was documented, in agreement with previous data ob-

tained under dry conditions on carbonates with a porosity of about

30% (e.g.,Baxevanis et al., 2006; Baud et al., 2009; Zhu et al., 2010;

Fig.12. a) Column diagram showing the density of twinned calcite crystals in the different samples. The data are grouped in two clusters to show the difference between the strain

localization areas and the rest of the sample; Microphotographs (cross-polarized nicols) and SEM images from sample 45 to bedding (b ec) and 0 to bedding (dee), the arrows point

out to some micro-cracks and the Tindicates the twinned crystals. For all the images the direction ofs1is vertical. (b) and (c) are taken within the CSB, (d) shows collapsed pores

and diffuse deformation, high density of twinned calcite is documented in picture (e).



15/19

Vajdova et al., 2012). Conversely,Cilona et al., (2012) documented

shear-enhanced dilation at 5 MPa ofPeff.under wet conditions and

strain rates of 107 s1, suggesting the occurrence of chemical

processes such as subcritical crack growth (seeBrantut et al., 2013

for a recent review).

Fig. 13compiles the yield points data in the stress space (PeQ)

and compares them with dry date fromVajdova et al.'s (2004)on

Indiana limestone (f 16%) and with Baud et al.'s (2009) on

Orfento and S. Maximin grainstones (f30 and 37%, respectively).

While the yielding strength of the carbonates from literature is

inversely proportional to porosity and grain size, Bolognano

grainstone turn out to be much stronger than one would predict by

considering only its grain size and porosity values. Since more

variability in porosity and grain size is present in these carbonates,

predicting the plastic yield envelops and the conditions for strain

localization, is not as reliable as it is in sandstones (Shultz et al.,

2010and references therein). For example, the strength of Bolog-

nano grainstones is higher than that of Orfento limestone (Baud

et al., 2009) and may be a consequence of a more abundant pres-

ence of echinoid fragments (Rustichelli et al., 2013). Indeed echi-

noid fragments promote the growth of syntaxial cement, which is

responsible of a strengthening of the grain contacts.

During the deformation experiments, both low- and high-angleto s1 deformation bands developed. Moreover, in agreement with

carbonates (Baud et al. 2009; Cilona et al., 2012) and sandstones

(e.g., Baud et al., 2004; Fortin et al., 2006; Wong and Baud, 2012 )

studies, a general positive correlation was found between the

conning pressure and the angle formed by DBs and the maximum

compression axis (Fig. 5). These latter results are consistent with

the statement that to form high-angle to s1 deformation bands

higher conning pressures and lower differential stresses are

needed with respect to low angle ones (i.e., Baud et al., 2004;

Fossen et al., 2011; Cilona et al., 2012).

6.2. Microstructural characterization of deformation bands

A systematic microstructural analysis was carried out on sam-ples deformed at different stress and strain conditions. For the

analyses, thin section observations and X-ray mCT data were

integrated. For comparison, the analyses were performed on host

rock, experimentally, and naturally deformed samples.

InFig. 2(a) we showed that the host rock contains alternating

laminae characterized by different values of sorting (Fig. 2a) with

porosity contrasts ofz4%. These contrasts are likely determined by

the variable concentration of bryozoans fossils, which are rich in

intragranular macro-pores (Rustichelli et al., 2013). A few percen-

tiles difference in porosity may strongly affect the mechanical

behaviour of a layer with respect to another in carbonates ( Cilona

et al., 2012) as well as in other rocks (e.g., Baud et al., 2012). Also

Louis et al. (2009) documented up to z8% of porosity difference

between low- and high-porosity layers in the Rothbach sandstones.

The crack surface area values we calculated for the undeformed

sample are one order of magnitude higher than those documented

by Wu et al. (2000) in their unstressed sample of Darley Dale

sandstone. The values measured on Bolognano are so high because

of the contribution of 10 s ofmm-long cracks emanating from pores

(Figs. 6 and 7). Indeed, in addition to the lower strength of calcite

with respect to quartz, the abundant intragranular pores in

Bolognano might have promoted the formation of pore-emanating

cracks during the natural deformation phases experienced by the

rock (Fig. 6aeb; Vajdova et al., 2004; 2010). Despite we docu-

mented high values of crack surface area also for the naturallydeformed sample, we highlight that the density within the

compaction band was similar to that in the surrounding host rock

(Fig. 8). This result suggests that within natural compaction bands

pressure solution prevails respect to grain crushing, in agreement

withTondi et al. (2006)and Cilona et al. (2012).

We compared the crack-density area values measured in this

study with those obtained by previous author in samples deformed

at similar levels of plastic strain. Our results were two-to-three

times lower than those published for porous carbonates (Vajdova

et al., 2010; Cilona et al., 2012), and up to a factor six higher than

those measured in sandstones (Wu et al., 2000).

Unlike other studies (Wu et al., 2000;Vajdova et al., 2010), we

did not aim to describe how the damage increases with the axial

strain; thus we did not calculate the Sv at different stages ofdeformation and constant Pc. Despite the different levels of plastic

strain experienced by the samples, a negative trend between the

conning pressure and Sv can be observed in Fig. 8(a). We docu-

mentedthe highest valuesofSv in the sample deformedat 5 MPa Pc

and the lowest (factor three less) in the sample deformed at 35 MPa

Pc. This difference might be causedby an interplayof grain crushing

and pore collapse: in the ductile regime the latter process is

dominant. After comparing the samples deformed at 25 MPa Pc

which experienced the same amount of plastic strain, we docu-

mented higher Sv values in the two samples where strain locali-

zation occurred. In the host rock, the 2D porosity values

underestimated the 3D ones by a factor of z1.5; Rustichelli et al.

(2012)documented similar ratios. The 2D porosity within the lab-

oratory compaction band underestimated the 3D one by a factorthree-to-four. The underestimation is higher within the DBs

because of the smaller pore size: image analysis accuracy is

strongly related to the spatial and contrast resolution of the

analyzed pictures (Fig. 9). Respect to the host rock, the porosity

decreased up-to-one fourth within the laboratory deformation

bands and up-to-one fth in the natural CB. Similar porosity re-

ductions are documented byCilona et al. (2012).

6.3. Rock heterogeneity and compaction localization

The specimens deformed at 25 MPa Pc showed ductile behav-

iour. Although in these experiments the recorded peaks were

almost equal, the bed-perpendicular specimen had the lowest

strength whereas, the bed-oblique specimen showed the highest

Fig. 13. Compilation of yield points of different porous carbonates: Bolognano (this

study), Indiana (Vajdova et al., 2004); Orfento Majella Limestone and St. Maximin

(Baud et al., 2009).



16/19

strength. Lower strength of the bed-perpendicular samples with

respect to the bed-parallel ones was also measured byBaud et al.

(2012) for the Diemelstadt sandstone. Conversely, Louis et al.

(2009) documented the highest strength in the bed-

perpendicular specimens and the lowest in the bed-parallel ones.

These latter authors claim that the distribution of grain contacts

controls the strength of the Rothbach sandstone. Our experimental

results are consistent with Bolognano grainstones promoting strain

localization within laminae sub-parallel to bedding characterized

by higher porosity and better sorting (Fig. 2). Additionally to

porosity, grain sorting plays a fundamental role on strain localiza-

tion: better-sorted layers are more keen to develop compaction

bands (Fossen et al., 2011; Cheung et al., 2012; Rustichelli et al.,

2012; Skurtveit et al., 2013).

The microstructures we documented varied with the orienta-

tions of sample with respect to bedding (Fig.11). These results are

similar to those presented by Louis et al. (2007) for Rothbach

sandstone, however the same authors demonstrated that one of the

45-to bedding samples, deformed at the conning pressure at

which CBs were predicted, did not develop compactive shear bands

but short and discontinuous compaction bands conned within

porous layers (Louis et al., 2009). This latter event likely did not

happen for the 45-to bedding sample of Bolognano because amacroscopic shear offset was detected on that specimen.

Despite few examples of bed-parallel compactive shear bands

are documented in some of the cross-beds of Bolognano Fm.

(Fig. 13a; Agosta et al., 2009), we suggest that during its burial

history the studied rocks unlikely experienced a differential stress

of magnitude equal to that one shown in Fig. 10(with vertical s1).

Conversely we cannot exclude that similar stress conditions were

reached during Middle-to-Late Pliocene when Majella Anticline

formed (Ghisetti and Vezzani, 2002). This scenario would imply

diffuse damage caused by the horizontal s1 and would justify the

high crack density measured in the host rock sample (Fig. 6aeb and

8).

We documented the highest twinning density in the 45-to

bedding sample, this value is justied by the occurrence of shearfailure; indeed calcite requires low shear stresses to initiate me-

chanical twinning and dislocation (Vajdova et al., 2010, 2012;

Cilona et al., 2012). In this study, mechanical twinning was

mainly documented within the overgrowth syntaxial cement

(Rustichelli et al., 2013 for detailed information). We suggest that

sample-scale variations of echinoids fragments amounts affect the

density of twinning (Fig. 12) and can inuence the interplay be-

tween grains crushing and mechanical twinning from one sample

to another. Although Vajdova et al. (2012)describes the mechanical

effects of the predominance of one or the other process, further

systematic analyses should be performed to address this issue.

In this paper we propose a scenario for compaction localization

in carbonates that is very different from what previously presented

for sandstones, for which rock homogeneity appears to promotethe development of compaction bands (Wang et al., 2008; Cheung

et al., 2012). Laboratory experiments on sandstones documented

CBs oriented perpendicular to s1, independently from the angle

formed by the specimens and the sedimentary bedding (e.g., Baud

et al., 2006; Fortin et al., 2006), the main effect of bedding in-

terfaces was to inhibit the propagation of compaction bands

through the sample (Louis et al., 2009) or increase their tortuosity

(Baud et al., 2012).

In limestones, to our knowledge, natural CBs have been mainly

documented parallel to bedding, onlyTondi et al. (2012)describes

bed-oblique CBs at the tips of well-developed faults. Bed-parallel

compaction bands tend to localize within the most porous and/or

coarser layers (Tondi et al., 2006; Agosta et al., 2009; Cilona et al.,

2012; Rustichelli et al., 2012; Tondi et al., 2012), and pre-existing

mechanical interfaces (e.g., carbonate hardground or bedding)

may also promote compaction localization (Cilona et al., 2011;

Rustichelli et al., 2012). These eld observations provide a

possible clue to explain why in previous laboratory studies, per-

formed on homogeneous porous carbonates, pure compaction

bands did not nucleate (seeWong and Baud, 2012for a full review).

Many systematic experimental studies documented either diffuse

deformation (Vajdova et al., 2012) or high-angle to s1 compactive

shear bands/shear-enhanced compaction bands (Baud et al., 2009;

Vajdova et al., 2010) associated to ductile failure. We suggest that

the absence of strong bedding-parallel heterogeneity inhibited the

systematic development of compaction bands in most of the

laboratory-deformed carbonates. Cilona et al. (2012) deformed

different strata of the Orfento Fm. from Madonna della Mazza

Quarry of the Majella Mountain, Italy (Tondi et al., 2006) and were

able to produce experimentally compaction bands in some of the

deformed strata. Our results are consistent with the observations of

Dautriat et al. (2011a)and conrmed the hypothesis ofCilona et al.

(2012): rock heterogeneity may enhance compaction localization.

A stress-independent kinematics of natural deformation bands

in the studied carbonates should be considered. Bolognano grain-

stones are characterized by alternations of horizontal and crossed

beds (Fig. 13;Rustichelli et al., 2013).Agosta et al. (2009)describedcompactive shear bands parallel to cross beds adjacent to bed-

parallel compaction bands localized within horizontal beds

(Rustichelli et al., 2012). Based upon our experimental data, and

comparing the orientations and kinematics of the deformation

bands developed in bed-oblique and bed-perpendicular samples,

we propose that under the same stress conditions (i.e. mean and

differential stress;Fig. 10) both compaction bands and bed-parallel

compactive shear bands can nucleate at the same time in strata

oriented oblique or perpendicular s1. It actually means that rock

heterogeneity can be responsible of the switching from a purely

volumetric deformation to a shear/volumetric one (Fig. 14b).

The strain localization into compaction bands or compactive

shear bands would then cause different types of stress perturbation

at the tips of these structures with implications on orientation anddistribution of secondary dilatant tail structures (e.g., dilation

bands, joints). From the prospective of reservoir exploration and

production it can be postulated that, due to the sedimentary ar-

chitecture, some portions of a reservoir could increase their con-

nectivity if a higher effective stress would cause the formation of

compactive shear band.

7. Conclusions

The presented study integrated eld and laboratory approaches

to investigate the effects of rock heterogeneity on compaction

localization in porous carbonates. A systematic set of triaxial

compaction experiments was performed on the Oligocene-

Miocene skeletal grainstones of Bolognano Fm., central Italy. Thecarbonates displayed shear-enhanced compaction and strain

hardening associated with various patterns of strain localization.

The brittle/ductile transition occurred at 12.5 MPa conning pres-

sure, and discrete compaction bands nucleated at 25 MPa. A posi-

tive correlation between conning pressure magnitude and the

angular value formed by individual deformation band and the

major principal stress axis was observed.

As natural compaction bands also the laboratory ones localized

within bryozoan-rich laminae, because of the z4% higher porosity

and better sorting.

We compared internal structure as well as the micromechanism

of laboratory compaction bands to those of natural one. Despite the

internal texture appeared to be very similar for both structures, in

natural compaction bands pressure solution dominates with



17/19


18/19

Besuelle, P., 2001. Compacting and dilating shear bands in porous rock: theoreticaland experimental conditions. J. Geophys. Res. 106 (B7), 13435e13442. http://dx.doi.org/10.1029/2001JB900011.

Blunt, M.J., Bijeljic, B., Dong, H., Gharbi, O., Iglauer, S., Mostaghimi, P., Paluszny, A.,Pentland, C., 2013. Pore-scale imaging and modelling. Adv. Water Resour. 51,197e216. http://dx.doi.org/10.1016/j.advwatres.2012.03.003.

Brantut, N., Heap, M., Meredith, P.G., Baud, P., 2013. Time-dependent cracking andbrittle creep in crustal rocks: a review. J. Struct. Geol. 52, 17 e43.

Brun, F., Mancini, L., Kasae, P., Favretto, S., Dreossi, D., Tromba, G., 2010. Pore3D: asoftware library for quantitative analysis of porous media. Nucl. Instrum.

Methods Phys. Res. Sect. A 615, 326e

332. http://dx.doi.org/10.1016/j.nima.2010.02.063.

Cheung, C.S.N., Baud, P., Wong, T., 2012. Effect of grain size distribution on thedevelopment of compaction localization in porous sandstone. Geophys. Res.Lett. 39, L21302. http://dx.doi.org/10.1029/2012GL053739.

Cilona, A., Agosta, F., Tondi, E., 2011. Preliminary results of the microstructural andtextural characterization of compaction bands affecting four porous carbonateformations. Geophys. Res. Abstr. 13, egu2011e10377. EGU General Assembly2011.

Cilona, A., Baud, P., Tondi, E., Agosta, F., Vinciguerra, S., Rustichelli, A., Spiers, C.J.,2012. Deformation bands in porous carbonate grainstones:eld and laboratoryobservations. J. Struct. Geol. 45, 137e157. http://dx.doi.org/10.1016/j.jsg.2012.04.012.

Cloetens, P., Barrett, R., Baruchel, J., Guigay, J.P., Schlenker, M., 1996. Phase objects insynchrotron radiation hard x-ray imaging. J. Phys. D Appl. Phys. 29, 133e146.http://dx.doi.org/10.1088/0022-3727/29/1/023 .

Cloetens, P., Pateyron-Salome, M., Bufere, J.Y., Peix, G., Baruchel, J., Peyrin, F.,Schlenker, M., 1997. Observation of microstructure and damage in materials byphase sensitive radiography and tomography. J. Appl. Phys. 81, 5878e5886.

Crescenti, U., Crostella, A., Donzelli, G., Raf, G., 1969. Stratigraa Della Serie Cal-carea Dal Lias al Miocene Nella Regione, vol. 8. Memorie della Societ aGeologicaItaliana, Marchigiano-Abruzzese, pp. 343e420.

Croize, D., Bjrlykke, K., Jahren, J., Renard, F., 2010a. Experimental mechanical andchemical compaction of carbonate sand. J. Geophys. Res. 115, B11204.http://dx.doi.org/10.1029/2010JB007697.

Croize, D., Ehrenberg, S.N., Bjrlykke, K., Renard, F., Jahren, J., 2010b. Petrophysicalproperties of bioclastic platform carbonates: implications for porosity controlsduring burial. Mar. Petrol. Geol. 27, 1765e1774.

Curran, J.H., Carroll, M.M., 1979. Shear stress enhancement of void compaction.J. Geophys. Res. 84, 1105e1112.

Cuss, R.J., Rutter, E.H., Holloway, R.F., 2003. The application of critical state soilmechanics to the mechanical behaviour of porous sandstones. Int. J. Rock Mech.Min. Sci. 40 (6), 847e862. http://dx.doi.org/10.1016/S1365-1609(03)00053-4.

Dautriat, J., Bornert, M., Gland, N., Dimanov, A., Raphanel, J., 2011a. Localizeddeformation induced by heterogeneities in porous carbonate analysed by multi-scale digital image correlation. Tectonophysics 503 (1), 100e116.

Dautriat, J., Gland, N., Dimanov, A., Raphanel, J., 2011b. Hydromechanical behavior

of heterogeneous carbonate rock under proportional triaxial loadings.J. Geophys. Res. 116, B01205. http://dx.doi.org/10.1029/2009JB000830.Deng, S., Aydin, A., 2012. Distribution of compaction bands in 3D in an aeolian

sandstone: the role of cross-bed orientation. Tectonophysics 574e575,204e218. http://dx.doi.org/10.1016/j.tecto.2012.08.037.

Dowd, B.A., Graham, H., Campbell, R.B., Marr, V.V., Nagarkar, S.V., Tipnis, L.A.,Siddons, D.P., 1999. Developments in synchrotron X-ray computed micro-tomography at the National Synchrotron Light Source. In: SPIE's InternationalSymposium on Optical Science, Engineering, and Instrumentation. InternationalSociety for Optics and Photonics, 1999, pp. 224e236.

Eichhubl, P., Hooker, J.N., Laubach, S.E., 2010. Pure and shear-enhanced compactionbands in Aztec Sandstone. J. Struct. Geol. 32, 1873e1886.

Faulkner, D.R., Jackson, C.A.L., Lunn, R.J., Schlische, R.W., Shipton, Z.K.,Wibberley, C.A.J., Withjack, M.O., 2010a. A review of recent developmentsconcerning the structure, mechanics and uid ow properties of fault zones.

J. Struct. Geol. 32, 1557e1575. http://dx.doi.org/10.1016/j.jsg.2010.06.009.Faulkner, D.R., Pett, G., Armitage, P.J., 2010b. A New High Pressure e High Tem-

perature Rock Deformation Apparatus. Gordon conference.Fortin, J., Stanchits, S., Dresen, G., Gueguen, Y., 2006. Acoustic emission and ve-

locities associated with the formation of compaction bands in sandstone.J. Geophys. Res. 111, B10203. http://dx.doi.org/10.1029/2005JB003854.

Fortin, J., Schubnel, A., Gueguen, Y., 2005. Elastic wave velocities and permeabilityevolution during compaction of Bleurswiller sandstone. Int. J. Rock Mechan.Min. Sci. 42, 873e889.

Fossen, H., Schultz, R.A., Torabi, A., 2011. Conditions and implications for compac-tion band formation in the Navajo Sandstone, Utah. J. Struct. Geol. 33,1477e1490. http://dx.doi.org/10.1016/j.jsg.2011.08.001.

Fossen, H., Schultz, R.A., Shipton, Z.K., Mair, K., 2007. Deformation bands in sand-stone: a review. J. Geol. Soc. 164, 755e769. http://dx.doi.org/10.1144/0016-76492006-036.

Ghisetti, F., Vezzani, L., 1998. Geometrie deformative. In: evoluzione cinematicadellAppennino centrale (Ed.), Studi Geologici Camerti 14, 127e154.

Ghisetti, F., Vezzani, L., 2002. Normal faulting, extension and uplift in the outerthrust belt of the central Apennines (Italy): role of the Caramanico fault. BasinRes. 14, 225e236.

Homand, S., Shao, J.F., 2000. Mechanical behaviour of a porous chalk and effect ofsaturating uid. Mech. Cohesive-frictional Mater. 5, 583e606.http://dx.doi.org/10.1002/1099 1484(200010)5:73.0.CO;2-J.

Issen, K.A., Rudnicki, J.W., 2000. Conditions for compaction bands in porous rock.J. Geophys. Res. 105, 21,529e21,536.

Jaeger, J.C., Cook, N.G.W., Zimmerman, R., 2006. Fundamentals of Rock Mechanics.Blackwell publishing.

Ji, Y., Baud, P., Vajdova, V., Wong, T.-F., 2012. Characterization of pore geometry ofIndiana limestone in relation to mechanical compaction. Oil Gas. Sci. Technol.67 (5), 753e775.

Kak, A.C., Slaney, M., 1988. Principles of Computerized Tomographic Imaging. IEEEPress.

Liteanu, E., Spiers, C.J., de Bresser, J.H.P., 2013. The inuence of water and super-

critical CO2 on the failure behavior of chalk. Tectonophysics 599, 157e

169.Louis, L., Baud, P., Wong, T.-f., 2009. Microstructural inhomogeneity and mechanical

anisotropy associated with bedding in Rothbach Sandstone. Pure Appl. Geophys166, 1063e1087. http://dx.doi.org/10.1007/s00024-009-0486-1.

Louis, L., Wong, T.-f., Baud, P., 2007. Imaging strain localization by X-ray radiographyand digital image correlation: deformation bands in Rothbach sandstone.

J. Struct. Geol. 29, 129e140.Louis, L., Wong, T-f., Baud, P., Tembe, S., 2006. Imaging strain localization by X-ray

computed tomography: discrete compaction bands in Diemelstadt sandstone.J. Struct. Geol. 28, 762e775. http://dx.doi.org/10.1016/j.jsg.2006.02.006.

Mair, K., Main, I., Elphick, S., 2000. Sequential growth of deformation bands in thelaboratory. J. Struct. Geol. 22, 25e42.

Mancini, L., Reiner, E., Cloetens, P., Gastaldi, J., Hartwig, J., Schlenker, M., Baruchel, J.,1998. Investigation of defects in AlPdMn icosahedral quasicrystals by combinedsynchrotron X-ray topography and phase radiography. Philos. Mag. A 78,1175e1194.

Marchegiani, L., Van Dijk, J.P., Gillespie, P.A., Tondi, E., Cello, G., 2006. Scalingproperties of the dimensional and spatial characteristics of fault and fracturesystems in the Majella Mountain, central Italy. In: Cello, G., Malamud, B. (Eds.),Fractal Analysis for Natural Hazards, Geological Society of London, SpecialPublication, vol. 261, pp. 113e131. http://dx.doi.org/10.1144/GSL.SP.2006.261.01.09.

Mollema, P.N., Antonellini, M.A., 1996. Compaction bands: a structural analog foranti-mode I cracks in aeolian sandstone. Tectonophysics 267, 209e228.http://dx.doi.org/10.1016/S0040-1951(96http://dx.doi.org/10.1016/S0040-1951(96)00098-4.

Olsson, W.A.,1999. Theoretical and experimental investigation of compaction bandsin porous rock. J. Geophys. Res. 104, 7219e7228. http://dx.doi.org/10.1029/1998JB900120.

Paterson, M.S., Wong, T-f., 2005. Experimental Rock Deformation; the Brittle Field,second ed. Springer-Verlag, Berlin, Federal Republic of Germany.

Pomar, L., Brandano, M., Westphal, H., 2004. Miocene tropical foramol-rhodalgal-bryomol associations of the western Mediterranean e a critical review. Sedi-mentology 51, 627e651. http://dx.doi.org/10.1111/j.1365-3091.2004.00640.x.

Risnes, R., Madland, M.V., Hole, M., Kwabiah, N.K., 2005. Water weakening of chalke mechanical effects of watereglycol mixtures. J. Petrol. Sci. Eng. 48 (1e2),21e36.

Rustichelli, A., Tondi, E., Agosta, F., Cilona, A., Giorgioni, M., 2012. Development anddistribution of bed-parallel compaction bands and pressure solution seams inthe Bolognano Formation carbonates (Majella Mountain, Italy). J. Struct. Geol.37, 181e199. http://dx.doi.org/10.1016/j.jsg.2012.01.007.

Rustichelli, A., Tondi, E., Agosta, F., Di Celma, C., Giorgioni, M., 2013. Sedimentologicand diagenetic controls on pore-network characteristics of OligoceneeMioceneramp carbonates (Majella Mountain, central Italy). AAPG Bull. 97 (3), 487 e524.http://dx.doi.org/10.1306/07311212071.

Schultz, R.A., Okubo, C.H., Fossen, H., 2010. Porosity and grain size controls oncompaction band formation in Jurassic Navajo Sandstone. Geophys. Res. Lett. 37,L22306. http://dx.doi.org/10.1029/2010GL044909.

Scisciani, V., Tavarnelli, E., Calamita, F., 2002. The interaction of extensional andcontractional deformation in the outer zones of the Central Apennines, Italy.

J. Struct. Geol. 24, 1647e1658.http://dx.doi.org/10.1016/S0191-8141(01)00164-X.

Skurtveit, E., Torabi, A., Gabrielsen, R.H., Zoback, M.D., 2013. Experimental investi-gation of deformation mechanisms during shear-enhanced compaction inpoorly lithied sandstone and sand. J. Geophys. Res. Solid Earth 118,4083e4100. http://dx.doi.org/10.1002/jgrb.50342.

Sok, R.M., Varslot, T., Ghous, A., Latham, S., Sheppard, A.P., Knackstedt, M.A., 2009.Pore scale characterization of carbonates at multiple scales: Integration ofMicroCT, BSEM and FIBSEM. In: Proceedings of the International Symposium ofthe Society of Core Analysts, Noordwijk, The Netherlands.

Tondi, E., Antonellini, M., Aydin, A., Marchegiani, L., Cello, G., 2006. The role ofdeformation bands, stylolites and sheared stylolites in fault development incarbonate grainstones of Majella Mountain, Italy. J. Struct. Geol. 28, 376e391.

Tondi, E., Cilona, A., Agosta, F., Aydin, A., Rustichelli, A., Renda, P., Giunta, G., 2012.Growth processes, dimensional parameters and scaling relationships of twoconjugate sets of compactive shear bands in porous carbonate grainstones,Favignana Island, Italy. J. Struct. Geol. 37, 53e64. http://dx.doi.org/10.1016/

j.jsg.2012.02.003.Tromba, G., Longo, R., Abrami, A., Arfelli, F., Astolfo, A., Bregant, P., Brun, F.,

Casarin, K., Chenda, V., Dreossi, D., Hola, M., Kaiser, J., Mancini, L., Menk, R.H.,Quai, E., Quaia, E., Rigon, L., Rokvic, T., Sodini, N., Sanabor, D., Schultke, E.,Tonutti, M., Vascotto, A., Zanconati, F., Cova, M., Castelli, E., 2010. In: The SYR-MEP Beamline of Elettra: Clinical Mammography and Bio-medical Applications.AIP Conference Proceedings, vol. 1266, pp. 18e23. http://dx.doi.org/10.1063/1.3478190.

http://dx.doi.org/10.1029/2001JB900011http://dx.doi.org/10.1029/2001JB900011http://dx.doi.org/10.1016/j.advwatres.2012.03.003http://refhub.elsevier.com/S0191-8141(14)00154-0/sref21http://refhub.elsevier.com/S0191-8141(14)00154-0/sref21http://refhub.elsevier.com/S0191-8141(14)00154-0/sref21http://dx.doi.org/10.1016/j.nima.2010.02.063http://dx.doi.org/10.1016/j.nima.2010.02.063http://dx.doi.org/10.1029/2012GL053739http://dx.doi.org/10.1029/2012GL053739http://refhub.elsevier.com/S0191-8141(14)00154-0/sref24http://refhub.elsevier.com/S0191-8141(14)00154-0/sref24http://refhub.elsevier.com/S0191-8141(14)00154-0/sref24http://refhub.elsevier.com/S0191-8141(14)00154-0/sref24http://refhub.elsevier.com/S0191-8141(14)00154-0/sref24http://refhub.elsevier.com/S0191-8141(14)00154-0/sref24http://dx.doi.org/10.1016/j.jsg.2012.04.012http://dx.doi.org/10.1016/j.jsg.2012.04.012http://dx.doi.org/10.1088/0022-3727/29/1/023http://refhub.elsevier.com/S0191-8141(14)00154-0/sref27http://refhub.elsevier.com/S0191-8141(14)00154-0/sref27http://refhub.elsevier.com/S0191-8141(14)00154-0/sref27http://refhub.elsevier.com/S0191-8141(14)00154-0/sref27http://refhub.elsevier.com/S0191-8141(14)00154-0/sref27http://refhub.elsevier.com/S0191-8141(14)00154-0/sref27http://refhub.elsevier.com/S0191-8141(14)00154-0/sref27http://refhub.elsevier.com/S0191-8141(14)00154-0/sref27http://refhub.elsevier.com/S0191-8141(14)00154-0/sref28http://refhub.elsevier.com/S0191-8141(14)00154-0/sref28http://refhub.elsevier.com/S0191-8141(14)00154-0/sref28http://refhub.elsevier.com/S0191-8141(14)00154-0/sref28http://refhub.elsevier.com/S0191-8141(14)00154-0/sref28http://refhub.elsevier.com/S0191-8141(14)00154-0/sref28http://refhub.elsevier.com/S0191-8141(14)00154-0/sref28http://refhub.elsevier.com/S0191-8141(14)00154-0/sref28http://refhub.elsevier.com/S0191-8141(14)00154-0/sref28http://dx.doi.org/10.1029/2010JB007697http://dx.doi.org/10.1029/2010JB007697http://dx.doi.org/10.1029/2010JB007697http://refhub.elsevier.com/S0191-8141(14)00154-0/sref30http://refhub.elsevier.com/S0191-8141(14)00154-0/sref30http://refhub.elsevier.com/S0191-8141(14)00154-0/sref30http://refhub.elsevier.com/S0191-8141(14)00154-0/sref30http://refhub.elsevier.com/S0191-8141(14)00154-0/sref30http://refhub.elsevier.com/S0191-8141(14)00154-0/sref30http://refhub.elsevier.com/S0191-8141(14)00154-0/sref30http://refhub.elsevier.com/S0191-8141(14)00154-0/sref31http://refhub.elsevier.com/S0191-8141(14)00154-0/sref31http://refhub.elsevier.com/S0191-8141(14)00154-0/sref31http://refhub.elsevier.com/S0191-8141(14)00154-0/sref31http://dx.doi.org/10.1016/S1365-1609(03)00053-4http://refhub.elsevier.com/S0191-8141(14)00154-0/sref33http://refhub.elsevier.com/S0191-8141(14)00154-0/sref33http://refhub.elsevier.com/S0191-8141(14)00154-0/sref33http://refhub.elsevier.com/S0191-8141(14)00154-0/sref33http://refhub.elsevier.com/S0191-8141(14)00154-0/sref33http://dx.doi.org/10.1029/2009JB000830http://dx.doi.org/10.1016/j.tecto.2012.08.037http://refhub.elsevier.com/S0191-8141(14)00154-0/sref38ahttp://refhub.elsevier.com/S0191-8141(14)00154-0/sref38ahttp://refhub.elsevier.com/S0191-8141(14)00154-0/sref38ahttp://refhub.elsevier.com/S0191-8141(14)00154-0/sref38ahttp://refhub.elsevier.com/S0191-8141(14)00154-0/sref38ahttp://refhub.elsevier.com/S0191-8141(14)00154-0/sref38ahttp://refhub.elsevier.com/S0191-8141(14)00154-0/sref39http://refhub.elsevier.com/S0191-8141(14)00154-0/sref39http://refhub.elsevier.com/S0191-8141(14)00154-0/sref39http://dx.doi.org/10.1016/j.jsg.2010.06.009http://dx.doi.org/10.1016/j.jsg.2010.06.009http://refhub.elsevier.com/S0191-8141(14)00154-0/sref41http://refhub.elsevier.com/S0191-8141(14)00154-0/sref41http://refhub.elsevier.com/S0191-8141(14)00154-0/sref41http://dx.doi.org/10.1029/2005JB003854http://refhub.elsevier.com/S0191-8141(14)00154-0/sref43http://refhub.elsevier.com/S0191-8141(14)00154-0/sref43http://refhub.elsevier.com/S0191-8141(14)00154-0/sref43http://refhub.elsevier.com/S0191-8141(14)00154-0/sref43http://refhub.elsevier.com/S0191-8141(14)00154-0/sref43http://dx.doi.org/10.1016/j.jsg.2011.08.001http://dx.doi.org/10.1016/j.jsg.2011.08.001http://dx.doi.org/10.1144/0016-76492006-036http://dx.doi.org/10.1144/0016-76492006-036http://refhub.elsevier.com/S0191-8141(14)00154-0/sref96http://refhub.elsevier.com/S019