11th Euroconference on rock physics and … 11th Euroconference on rock physics and geomechanics 'Holistic rock physics: integrating theory, observation and applications in space and

Abstracts

11th Euroconference on rock physics and geomechanics

'Holistic rock physics: integrating theory, observation and applications in space and time'

6–11 September 2015, Ambleside, UK

Eurocinference abstracts.indd 1 14/08/2015 13:28:03

2


3

11TH EURO CONFERENCE ON ROCK PHYSICS AND GEOMECHANICS

'Holistic rock physics: integrating theory, observation and applications in space and time'

6–11 September 2015, Ambleside, UK

It is a pleasure to welcome you to Ambleside for the 11th EURO Conference on Rock Physics and Geomechanics. Our overarching theme for this event is that of ‘holistic rock physics’ — the concept of approaching problems in a multifaceted way, gaining insight from experimental, field, and numerical analyses to develop a better understanding of the underlying mechanisms which we deal with.

This conference is an exceptional platform to bring the community together, to discuss each others progress, and to drive new innovations and collaborations. We have supported this aim by attempting to grant both oral and poster sessions the time necessary for productive discussion, and by ensuring that no sessions are held concurrently. There is also a social programme scheduled, and we hope you will join us all for the icebreaker on Sunday, the conference banquet on Tuesday, and the final night barbecue on Thursday.

The location here at the University of Cumbria in Ambleside grants us a fantastic opportunity to view some of the most impressive geology in the UK. The two fieldtrips will run concurrently on Wednesday afternoon. If you wish to join one of these trips and have not yet made a choice online, then please approach one of the conference organisers and we will try to fit you in. Unfortunately space is limited. If you are joining us on these trips please be on the buses by 13.30 at the latest.

We look forward to a productive and interesting meeting, and on behalf of the Organising and Scientific Committees, I thank you for your participation.

Philip Benson

Organising committeePhilip Benson, Pete Rowley, Sergio Vinciguerra, Ian Main, Yan Lavallée, Tom Mitchell, Dan Faulkner, Linda Hetherington.

Scientific Advisory CommitteeCino Viggiani, Philip Benson, Pierrre Bésuelle, Gary Couples, Christian David, Yves Guéguen, Patrick Baud.


4

Information for presenters

Oral presentationsPlease ensure you have either emailed your presentation to [email protected], or given your presentation to an organiser at least 3 hours before your presentation. We can run the presentation directly from your memory stick but we would like to check that all animations, fonts etc display properly on the system available.

All submitted talks are allotted 25 minutes, which includes time for questions. Please keep to these timings, as we have a very busy schedule. Session chairs will ask you to stop if you over-run.

Invited talks are 50 minutes including questions.

Poster presentations

All posters are displayed throughout the conference, and can be put up from Sunday afternoon. Each poster has been allocated a session and board number. Within the poster room the posters alternate between session one and session two (i.e. session 1 poster 4 is next to session 2 poster 4, etc). This is to ensure a good distribution of both posters and viewers on both days so people are not crowded in. Please have your posters removed by 6 pm on Thursday. Any posters remaining at this point will be recycled.

Sponsors

This conference, and in particular the numerous zero-fee student registrations, would not have been possible without the generous support of our sponsors. We are indebted to them for their generous help, and hope you can take the time during the conference to visit the stands of those who are displaying with us.

Gold sponsors

Silver sponsors

Other sponsors and supporters


5

Conference schedule

09.0

0

10.0

0

11.0

0

12.0

0

13.0

0

14.0

0

15.0

0

16.0

0

17.0

0

18.0

0

19.0

0

20.0

0

Sund

ay

6th

Mon

day

7th

Frid

ay11

thTu

esda

y8t

hTh

ursd

ay10

thW

edne

sday

9th

08.0

0

Ses

sion

1

Ses

sion

2

Ses

sion

4

Ses

sion

5

Ses

sion

6

Ses

sion

7

Pos

ter S

essi

on 1

Ses

sion

8

Ses

sion

9

Ses

sion

10

Ses

sion

11

Pos

ter S

essi

on 1

Pos

ter S

essi

on 2

Pos

ter S

essi

on 2

Lunc

hLu

nch

Lunc

h

Lunc

h

Exc

ursi

on

Load

bus

es fr

om 1

3.00

depa

rt at

13.

30

Ses

sion

3

Bus

es to

ban

quet

ven

ue

Cha

ir - N

icol

as B

rant

ut

Cha

ir - N

icol

as B

rant

ut

Wel

com

e - P

. Ben

son

Key

note

- C

. Spi

ers

Key

note

- P.

You

ngK

eyno

te -

I. Ja

ckso

nK

eyno

te -

E. R

utte

r

Pos

ter p

ico

talk

s

Iceb

reak

er

Reg

istra

tion

Ban

quet

Bar

becu

e

Dep

art

Pos

ter p

ico

talk

s

Bre

akB

reak

Bre

akB

reak

Bre

ak

Bre

akB

reak

Key

note

- B

. Eva

ns

Effe

cts

of fl

uids

on

faul

t fric

tion

and

heal

ing

beha

viou

r

The

britt

le-d

uctil

e tra

nsiti

on in

th

e E

arth

: A tr

ip to

the

zoo

Faul

ting,

cat

acla

stic

flow

an

d fri

ctio

n of

san

dsto

nes

Mod

els

and

mec

hani

cal

test

s: s

tudi

es o

f vis

coel

astic

an

d po

roel

astic

beh

avio

urLa

bora

tory

ear

thqu

akes

P. M

ered

ith -

A ho

listic

rock

ph

ysic

s ap

proa

ch.

Cha

ir - N

icol

as B

rant

ut

Cha

ir - G

eorg

Dre

sen

Cha

ir - G

eorg

Dre

sen

Cha

ir - P

hil B

enso

n

Cha

ir - P

hil B

enso

n

Cha

ir - P

ete

Row

ley

Cha

ir - P

ete

Row

ley

Cha

ir - T

om M

itche

ll

Cha

ir - T

om M

itche

ll


6

Low NookBuildingUniversity Information Point

EXIT

Charlotte Mason Building

To get to Charlotte Mason Building from the main car park, follow the dotted line and walk 400 metres up Nook Lane.

Directions

This car park is for visitor permits and disabled visitors only.

Notice

Parking for staff and student permit holders only.

Notice

Ready fromNovember

2013

Ready fromDecember

2013

ReadySpring 2014

ENTRANCETO CAMPUS

Car access for visitor permit holders and disabled. Max. width of lane is 180cms

1

2

3

5

6

7

8

9

4

H CJ

HA B

F

DE

Building access

Car park

Designated disabled persons parking bay

Building in use

Building not in use

Campus map key

University buildings noCharlotte Mason Building 1 – Gateway Reception 1 – ICT open access 1 – Library 1 – Video-conferencing suite 1Low Nook 4

Student residences Fairfi eld Hall C

H nyllevleHJ llefsnaW

University buildings noDrama studio 2Little Barn 3

5 telliMScale How 6Store (Sports hall) 7The Barn Theatre 8The Beehive 9

Student residences noAshfi eld ABeechfi eld BGreenbank North DGreenbank South E

F enedlezaHHeathfi eld G

Buildings in use

Buildings not in use

University of CumbriaAmbleside campus

AddressUniversity of Cumbria, Rydal Road, Ambleside,Cumbria LA22 9BB

Telephone015394 30274

Visitwww.cumbria.ac.uk

© U

niversity of C

umbria Sep

tember 20

13 (U

OC

060)


7

Low NookBuildingUniversity Information Point

EXIT

Charlotte Mason Building

To get to Charlotte Mason Building from the main car park, follow the dotted line and walk 400 metres up Nook Lane.

Directions

This car park is for visitor permits and disabled visitors only.

Notice

Parking for staff and student permit holders only.

Notice

Ready fromNovember

2013

Ready fromDecember

2013

ReadySpring 2014

ENTRANCETO CAMPUS

Car access for visitor permit holders and disabled. Max. width of lane is 180cms

1

2

3

5

6

7

8

9

4

H CJ

HA B

F

DE

Building access

Car park

Designated disabled persons parking bay

Building in use

Building not in use

Campus map key

University buildings noCharlotte Mason Building 1 – Gateway Reception 1 – ICT open access 1 – Library 1 – Video-conferencing suite 1Low Nook 4

Student residences Fairfi eld Hall C

H nyllevleHJ llefsnaW

University buildings noDrama studio 2Little Barn 3

5 telliMScale How 6Store (Sports hall) 7The Barn Theatre 8The Beehive 9

Student residences noAshfi eld ABeechfi eld BGreenbank North DGreenbank South E

F enedlezaHHeathfi eld G

Buildings in use

Buildings not in use

University of CumbriaAmbleside campus

AddressUniversity of Cumbria, Rydal Road, Ambleside,Cumbria LA22 9BB

Telephone015394 30274

Visitwww.cumbria.ac.uk

© U

niversity of C

umbria Sep

tember 20

13 (U

OC

060)


8

Fieldtrips

Trips take place on Wednesday, and buses should be loading at 13.00 for departure by 13.30. Late runners will be left behind. Please ensure you have signed up in advance for the trip you wish to attend if you have not already done so online. See one of the conference staff for more information.

Trip 1: Andesitic Sills and Volcani-clastic Sediments of the Borrowdale Volcanics Group at Honister Slate Mine, English Lake District.

Leader: Dr Hugh Tuffen, Lancaster Environment Centre

Volcanic activity in the Ordovician Lake District (~460 Ma) created the ~6 km-thick Borrowdale Volcanic Group (BVG) which is thought to record several overlapping and subsiding volcanic fields at an extensional continental margin, beneath which there was active subduction of oceanic lithosphere. This occurred in response to SE-directed subduction and closure of the Iapetus Ocean. At the head of the Buttermere valley a wide variety of volcanic rock types is exposed above the unconformable contact with the older Skiddaw Slates. This excursion will allow us to look at the ~300 m-thick Honister Slate Formation, to study andesitic sills thought to have intruded within thick water-lain volcani-clastic sediments, and to examine a complex sequence of volcanic breccias and tuffs that may record explosive magma-water interaction.

Trip 2: The Volcanic and Sedimentary Geology of Torver Common, Coniston, English Lake District.

Leader: Dr Ruth Siddall, University College London

The geology of Torver Common and the Old Man of Coniston expose the transition of a Caledonian active continental margin to a marine basin. A series of pyroclastic lithologies belonging to the Borrowdale Volcanic Group are overlain by a transgressive sequence of carbonate and clastic rocks of the Coniston Limestone Formation. These strata range in age from early Ordovician to latest Silurian. Tectonically, the rocks have been subjected to a transpressive cleavage-forming event during the Late Caledonian Acadian Orogeny. The area has been worked for slate since antiquity and this industry boomed in the early 19th Century, when the local green slate was exported worldwide. This field excursion will visit this stratigraphy on Torver Common, near Coniston, taking in the sedimentology and palaeontology of the key units.


9

TALKSDAY 1


10

Day 1 — Session 1 — Introduction

TIME-DEPENDENT DEFORMATION AND FAILURE: A HOLISTIC ROCK PHYSICS APPROACH

P Meredith

Rock & Ice Physics Laboratory, Department of Earth Sciences, University College London, London WC1E 6BT, UK.

Under upper crustal conditions most rocks accommodate deformation in a brittle manner through cracking, fracturing and faulting. Cracks can occur from the grain scale up to the crustal scale. Under all round compressive stress conditions, macroscopic fractures or faults result from the coalescence of numerous microscopic, grain-scale cracks. Deformation in the crust occurs over a wide range of strain rates, from the very slow rates associated with tectonic loading up to the very fast rates occurring during earthquake rupture. It is also now well-established that reactions between chemically-active pore fluids and the rock matrix can lead to time-dependent crack growth in rocks. In turn, this can allow them to deform and fail over extended periods of time at stresses well below their short-term strength, and even at constant stress.

The traditional way of representing rock deformation graphically is through the stress-strain curve which plots differential stress (1 — 3) against axial, radial or volumetric strain. Although extremely useful, this is inherently a bulk averaging procedure, and therefore provides only a global description of the failure process. So, in order to obtain a complete understanding of crustal dynamics we require a detailed knowledge of the time-dependent mechanical behaviour of rocks. Such knowledge should be based on micromechanics, but also provide an adequate macroscopic description. One way of moving towards such a full description has been to combine the traditional external, macroscopic measurements of rock mechanics with the internal, microscopic measurements of petrophysics to produce the holistic approach of integrated rock physics. The challenge is to establish a relationship between the internal, microstructural state of the rock and the macroscopically observable external quantities.

Here, we present a number of examples of attempts to reconcile these ideas through measurements of stress and strain evolution during deformation with simultaneous measurements of the evolution of other key rock physical properties such as elastic wave speeds, acoustic emission output, porosity and permeability. These parameters provide complementary information about the deformation process; for example, elastic wave speeds provide information about the overall state of internal crack damage, while acoustic emissions provide information about the contemporary rate of damage accumulation. Likewise, porosity describes the magnitude of the internal void (damage) space, whereas permeability describes how well that space is connected to provide percolation pathways. Overall, the combined data are able to explain both the complex shape of the stress-strain curve during constant strain rate loading and the shape of creep curves during constant stress loading, thus providing a unifying framework to describe the time-dependent mechanical behaviour of rock.


11

Day 1 — Session 1 — Keynote — 09.25

EFFECTS OF FLUIDS ON FAULT FRICTION AND HEALING BEHAVIOUR: INTEGRATING EXPERIMENTAL, MICROPHYSICAL AND OBSERVATIONAL

APPROACHES

C J Spiers

HPT Laboratory, Dept. of Earth Sciences, Faculty of Geosciences, Utrecht University, Netherlands

Understanding the effects of fluid-rock interaction on fault friction and healing/sealing behaviour is central not only to understanding natural seismogenesis and phenomena such as fluid trapping, but also to evaluating the risks of fault reactivation and induced seismicity posed by subsurface resources production and by geological storage of CO2. Microstructural studies on natural fault rocks deformed in the mid and upper crust, including those sampled in fault drilling projects, frequently show evidence for i) fluid-related reactions forming an anastomosing phyllosilicate network, ii) pressure solution and cataclasis of clast phases, and iii) dilatation and cementation of fractures, cracks and pores. Moreover, decades of friction and healing experiments on simulated granitic, gabroic, quartz and more recently phyllosilicate-quartz, carbonate and evaporite gouges, have shown that the presence of an aqueous pore fluid, or even water vapour, strongly influences the frictional behaviour of these materials. This has long been recognised to point to the operation of fluid-assisted deformation mechanisms, such as stress corrosion cracking or diffusive mass transfer (pressure solution). In particular, recent low velocity friction experiments performed on simulated carbonate, evaporite and quartz gouges, with varying amounts of phyllosilicate, indicate that water-assisted mass transfer processes are key to determining whether frictional slip is velocity-strengthening (stable) or velocity weakening (potentially seismogenic). Supercritical carbon dioxide, on the other hand, introduced under CO2 storage conditions, has little effect on the frictional behaviour of either dry or wet fault gouges. An important trend emerging from carbonate, evaporate and quartz rich gouges tested at sliding velocities below 100 µm/s, is a transition from velocity strengthening at low temperatures, to velocity weakening at intermediate temperatures, and back to velocity strengthening at high temperatures, delineating three regimes of steady state frictional behaviour. Where sample volume change has been measured or estimated, the velocity weakening regime is further characterised by significant porosity and permeability development. This behaviour and the healing/sealing behaviour that occurs when shearing is stopped are strongly influenced by water content. This all leads to the conclusion that a micromechanism-based description of the frictional behaviour of gouge-filled faults, under mid and upper crustal conditions, needs to account for diffusion and stress corrosion cracking mechanisms, and for both dilatant and non-dilatant slip on grain and phase boundaries. First attempts to do this, assuming diffusive mass transfer as the fluid-assisted deformation mechanism, successfully predict the three-regime behaviour seen in experiments, as well as other key observations, including evidence in carbonates for superplastic deformation processes operating in loacalized, mirror-like slip bands. Both steady state and transient frictional behaviour similar to that seen in experiments can be predicted. The key factor here controlling both frictional response (i.e a, b, a-b and Dc in the terminology of RSF modelling), porosity evolution and healing turns out to be competition between dilatation due to intergranular slip versus flow and compaction by diffusive mass transfer. In particular, velocity-weakening slip, hence seismic rupture nucleation, and 'interseismic' strength recovery are predicted to be caused by the effects of water in promoting compaction by diffusive transport during and after dilatant shear.


12

Day 1 — Session 1 — 10.15–10.40

PERMEABILITY ANISOTROPY AND FRACTURE HEALING IN SEDIMENTARY FORMATIONS IN A HYDRO-GEOTHERMAL CONTEXT

L Griffiths1, M J Heap1, F Wang1, D Daval2, H Albert Gilg3, P Baud1, A Genter4 and J Schmittbuhl1

1Institut de Physique de Globe de Strasbourg, Université de Strasbourg/EOST, CNRS UMR 7516, France2Laboratoire d’Hydrologie et de Géochimie de Strasbourg, Université de Strasbourg/EOST, CNRS UMR 7517, Strasbourg, France

3Munich Lehrstuhl für Ingenieurgeologie, Technische Universität München, Arcisstr. 21, 80333, Munich, Germany4ES-Géothermie, 3 Chemin du Gaz, Haguenau, 67500, France

Fluid circulation in geothermal reservoirs is greatly dependent on the geometry and hydraulic properties of fractures. The Soultz-sous-Forêts site located in the Upper Rhine Graben in Alsace, France, consists of a granitic reservoir overlain by a 1.4 km-thick sedimentary succession. Core analysis and borehole wall imagery collected from reconnaissance well EPS1, drilled to a depth of 2230 m, revealed an extensive fracture network throughout the granite and overlying sediments, including both open fractures and fractures filled through mineral precipitation (primarily quartz, barite, calcite, and galena). Here we present an experimental study that aims to provide insights into the impact of healed or partially-healed fractures on permeability anisotropy in the Triassic Buntsandstein sandstone (1000–1400 m depth). We targeted borehole samples that best represented the variability of fractures within the Buntsandstein unit for our study. Forty cylindrical core samples (40 mm in length and 20 mm in diameter) were prepared from the chosen borehole samples such that they contained healed or partially-healed fractures either parallel or perpendicular to their axis. We also prepared samples of the intact host rock. These samples were then subject to porosity and permeability measurements, and thin sections were made for Scanning Electron Microscopy (SEM) and Electron Backscatter Diffraction (EBSD) to characterise the nature of the fractures and the precipitated minerals. Permeability measurements of the Buntsandstein host rock yielded values ranging from 10–15 to 10–17 m2. Microscopic analysis suggests that prevalent pore-filling clays can explain the low permeability of the sandstone host rock. Additionally, we found that fractures may present a conduit for or a barrier to flow, depending on the extent of healing and the nature of the filling. It is likely that these fractures once represented effective conduits for flow but, as minerals were precipitated from the circulating hydrothermal fluids, the fractures healed and switched from providing conduits for flow to presenting barriers to flow. As a result, permeability anisotropy within these units likely changed over time. Ongoing efforts are focused on providing estimates for the time scale over which open fractures can heal. An improved knowledge of the permeability of fractures will provide a greater understanding of fluid circulation in the area and may have serious repercussions for the ongoing fluid flow modelling of the geothermal reservoir, where low permeability inhibits fluid convection and ultimately the transfer of heat through the system.


13

Day 1 — Session 1 — 10.40–11.05

MECHANICAL INSTABILITY INDUCED BY WATER WEAKENING IN LABORATORY FLUID INJECTION TESTS

C David1, J Dautriat2, J Sarout2, C Delle Piane2, B Menendez1, R Macault1,2 and D Bertauld1,2

1Department of Geosciences and Environment, University Cergy-Pontoise, France2CSIRO Energy, Perth, Western Australia

To assess water-weakening effects in reservoir rocks, previous experimental studies have focused on changes in the failure envelopes derived from mechanical tests conducted on rocks fully saturated either with water or with inert fluids. So far, little attention has been paid to the mechanical behavior during fluid injection under conditions similar to enhanced oil recovery operations. We studied the effect of fluid injection on the mechanical behavior of the weakly consolidated Sherwood sandstone in laboratory experiments. Our specimens were instrumented with 16 ultrasonic P wave transducers for both passive and active acoustic monitoring during loading and fluid injection to record the acoustic signature of fluid migration in the pore space and the development of damage. Calibration triaxial tests were conducted on three samples saturated with air, water, or oil. In a second series of experiments, water and inert oil were injected into samples critically loaded up to 80% or 70% of the dry or oil-saturated compressive strength, respectively, to assess the impact of fluid migration on mechanical strength and elastic properties. The fluids were injected with a low back pressure to minimize effective stress variations during injection. Our observations show that creep takes place with a much higher strain rate for water injection compared to oil injection. The most remarkable difference is that water injection in both dry and oil-saturated samples triggers mechanical instability (macroscopic failure) within half an hour whereas oil injection does not after several hours. The analysis of X-ray computed tomography images of post-mortem samples revealed that the mechanical instability was probably linked to loss of cohesion in the water-invaded region.


14

Day 1 — Session 2 — 11.30–11.55

FRICTIONAL INSTABILITIES AND MINERAL CARBONATION OF BASALTS TRIGGERED BY INJECTION OF PRESSURIZED H2O- AND CO2-RICH FLUIDS

P Giacomel1, E Spagnuolo2, A Marzoli1, M Nazzari2, N Youbi3,4 and G Di Toro1,2,5

1Dipartimento di Geoscienze, Università degli Studi di Padova, Padua, Italy2Istituto Nazionale di Geofisica e Vulcanologia, Rome, Italy

3Geology Department, Faculty of Sciences, Semlalia, Cadi Ayyad University, Prince Moulay Abdellah Boulevard, P.O. Box 2390, Marrakech, Morocco

4Centro de Geologia da Universidade de Lisboa (CeGUL), Faculdade de Ciências (FCUL), Departamento de Geologia (GeoFCUL), Campo Grande C6, Lisboa, Portugal

5School of Earth, Atmospheric & Environmental Sciences, University of Manchester, UK

Mineral carbonation of continental flood basalts by underground injection is a technique proposed for long-term storage of anthropogenic CO2. But the injection of pressurized fluids may result in induced seismicity. Here we discuss the effects of the injection of pure CO2, distilled H2O and H2O+CO2 fluid mixtures on basalt-built experimental faults. The experiments reproduced the stress conditions (normal stress <15 MPa) of the pilot projects for CO2 storage that exploit natural reservoirs located at few hundreds of meters depth. Since mineral carbonation requires the injection of H2O+CO2 fluid mixtures (or of pure CO2 in water reservoirs), the efficiency of the storage technique is limited by the low solubility of CO2 in H2O (30–70 g of CO2 per kg of H2O at the reservoir temperatures and depths).

We performed 10 friction experiments at room temperature on pre-cut basalts samples (50/30 mm external/internal diameter) with the rotary shear apparatus SHIVA. The basalts had a different degree of alteration determined with optical microscope, X-ray fluorescence and diffraction analysis. Samples were first pre-loaded to 15 MPa normal stress and to 5 MPa shear stress at an initial fluid pressure of 2.5 MPa; then, fluid pressure was increased of 0.1 MPa every 100 s to induce fault instability. The 'main' fault instability corresponded to the slip episode with slip rate V >0.3 m/s. The instability resulted from the unbalance between the imposed shear stress (5 MPa) and the time-dependent fault shear strength. The latter decreased with time because of (1) the step-increase of the fluid pressure and, (2) the chemical evolution of the sliding surface. In experiments performed with distilled H2O and H2O+CO2 fluids and, to a less extent, with pure CO2, the main instability was preceded by creep and short living slip bursts with 0.001 < V < 0.3 m/s. The main instability occurred at higher fluid pressures (1) in less altered basalts (regardless of the fluid composition) and, (2) in pure CO2 fluids given the same degree of alteration of the basalts.

Fluids collected after the experiments were analysed with ion-chromatography and the sliding surfaces investigated with Micro-Raman spectroscopy. The fluids recovered from the experiments with H2O+CO2 mixtures had a 2- and 5-fold increase of Mg2+ and Ca2+ concentration, respectively, with respect to those performed with distilled H2O. Only in the case of the H2O+CO2 fluid mixture experiments, the release of Mg2+ and Ca2+ resulted in the precipitation of dolomite and calcite grains (i.e., mineral carbonation).

Our experimental dataset suggests that fault instability is controlled both by the chemical composition of the fluid and the state of alteration of the basalt. The precipitation of calcite and dolomite in our short-in-duration experiments implies that mineral carbonation is an efficient mechanism for CO2 long-term underground storage. However, the low solubility of CO2 in water requires the injection of large quantities of fluid to store a significant amount of CO2, increasing the hazard associated to induced seismicity.


15

Day 1 — Session 2 — 11.55–12.20

THE FRICTIONAL PROPERTIES OF FAULTS AT SHALLOW DEPTHS: IMPLICATIONS FOR RUPTURE PROPAGATION

Nicola De Paola, Rachel Unwin, Rachael Bullock, Rosanne Murray, Mark Stillings and Robert E Holdsworth

Rock Mechanics Laboratory, Department of Earth Sciences, University of Durham (UK)

Most synoptic models of faults assume the presence of a shallow stable, velocity-strengthening aseismic region due to the presence of incohesive gouges, poorly lithified continental sediments (continental faults) and phyllosilicate-rich rocks (accretionary prisms at subduction zones). The near-surface portions of faults are therefore viewed as effective energy sinks with the potential to arrest/slow down the propagation of earthquakes, preventing them from reaching the surface. However, recent events, such as the 2009 Mw 6.3 L’Aquila and 2011 Mw 9.0 Tohoku-Oki earthquakes, have demonstrated that moderate/large co-seismic ruptures can propagate to the surface causing vast damage and destructive tsunamis.

In order to better understand rupture propagation at shallow depths, we investigated the frictional properties of a range of bedrock lithologies, typical of the oceanic (gabbros) and continental crusts (granite, limestone), together with phyllosilicate-bearing lithologies typical of subduction zones and continental sedimentary deposits. Laboratory experiments have been performed in a low to high velocity rotary shear apparatus, on granular materials with grainsize up to 200 µm, under dry, water- and brine-saturated conditions, at slip rates ranging from 10 µm/s up to 1 m/s, with normal loads up to 18 MPa and displacements up to 1 m.

Velocity step experiments performed at sub-seismic slip rates (10–100 µm/s) on dry, water- and brine-saturated granite and calcite rocks show that velocity strengthening behaviour evolves to velocity-neutral/-weakening behaviour due to slip localization attained after critical displacements of a few tens to hundreds of mm. The critical displacement value is inversely proportional to the applied normal load. Dry, water- and brine-saturated gabbros show velocity-weakening behaviour and slip localization regardless of the displacement attained and applied normal load. Dry, water- and brine-saturated phyllosilicate-rich gouges show velocity-strengthening behaviour for any applied displacements and normal loads. Dry continental sediments show initial velocity-strengthening behaviour evolving to velocity-neutral/-weakening behaviour for increasing displacement. At low normal loads (1 MPa), water-saturated continental sediments show velocity-strengthening behaviour, regardless of the displacement attained, and an evolution to velocity-weakening behaviour with increasing displacement at higher normal loads (18 MPa). Conversely, brine-saturated continental deposits always show velocity-strengthening behaviour, regardless of the displacement attained and applied normal load. Cyclic slide-hold-slide experiments show that, after sliding at sub-seismic slip rates, static friction increases with time according to a logarithmic relationship (fault healing) for almost all tested materials under dry, water- and brine-saturated conditions. The only exceptions are organic-rich oil-mature black shales, which show a decrease in static fault friction with time (negative fault healing). The positive and negative healing rates tend to increase under water- and brine-saturated conditions, respectively.

Our experimental results suggest that the frictional stability and healing rates of shallow fault patches will be controlled by factors such as depth/thickness (normal load), mineralogical composition, organic matter content, presence/composition of fluids, displacement and slip history (slip localization). Our findings may explain many of the discrepancies observed between the behavior of real earthquakes and that predicted by fault zone models, which rely on the simplistic assumption of a uniform, stable sliding frictional behavior at shallow depths.


16

Day 1 — Session 2 — 12.20–12.45

DYNAMIC WEAKENING AND FRACTURE ENERGY IN EXPERIMENTS AT SEISMIC CONDITIONS

Stefan Bjorklund Nielsen1, Elena Spagnuolo2, Marie Violay4 and Giulio Di Toro3

1Earth Sciences, Durham University, United Kingdom. 2Sismologia e Tettonofisica, INGV, Roma, Italy.

3School of Earth, Atmospheric & Environmental Sciences, University of Manchester, United Kingdom.4Rock Mechanics Laboratory, EPFL, Lausanne, Switzerland.

Fault rocks undergo abrupt dynamic weakening and lubrication during seismic slip, reputedly associated to thermally triggered physico-chemical processes. Recent experiments systematically explore rock friction under crustal earthquake conditions (slip rate 1–6 m/s, normal stress 5–50 MPa, water pressure or dry, various lithologies). The evolution during experiments is confronted to various thermal weakening models: flash weakening (FW), superplastic diffusion creep, melt lubrication (ML), fluid pressurization (FP). Our findings are that in the absence of melting and/or pressurization (e.g., dry carbonatic rocks) the whole frictional hysteresis cycle is explained by FW of contact asperities, provided that the interface temperature is accurately computed including the effect of heat sinks (latent heat of phase transitions), in a revised version of the FW model proposed by Archard (1958) and Rice (2006). In silicatic rocks under coseismic conditions, initial flash-weakening is followed by melting and subsequently behaves compatibly with the ML of Nielsen et al. (2008, 2010). The effects of water pore pressure on the mechanical evolution vary subtly depending on lithology and amount of sliding. However, experiments performed in the presence of water seem to indicate that thermal pressurization is not the dominant dynamic weakening mechanism as other thermal weakening act earlier and more efficiently. Frictional work in excess of minimum dynamic level, obtained in a number of experiment, yield values comparable to estimates from earthquake data for slips up to about u=1cm (~104 J/m2), to increase gradually with slip up to about 106 J/m2. However, it appears that earthquake fracture energy estimates are is slightly larger than the laboratory measures in the range of slip 0.1<u<10.


17

Day 1 — Session 3 — 14.00–14.25

FRICTION AND WEAR

Z Reches1, Y Boneh1,2, X Chen1,3, A Sagy4 and V Lyakhovsky4

1School of Geology and Geophysics, University of Oklahoma, Norman, OK, USA2Currently at: Earth and Planetary Sciences, Washington University, St. Louis, MO, USA

3Currently at: School of Geosciences, UT at Austin, TX, USA4Geological Survey of Israel, Jerusalem, Israel

Rabinowich stated in 1955: “When contacting surfaces are rubbed together, two phenomena, namely friction and wear, invariably accompany the sliding process.…[and] it may be said that high friction and high wear go together. However, the friction and wear processes are known to be very complex, and it is not surprising that no simple relationship…. has been found.” This statement is still valid, and we present an attempt to resolve part of these relations in fault-zones.

Our analysis is based on an extensive series of shear experiments at moderate to high slip-velocity (0.001–1.0 m/s), and low to moderate normal stress (0.5–7 MPa) that were conducted on unconfined, solid rock blocks (granite, tonalite, diorite, dolomite, limestone, and quartzite) and on confined rock powders (talc, calcite, dolomite, and powder mixtures). Monitoring stresses, temperature, fault-normal velocity, and CO2/H2O emission allowed to determine the friction-wear relations listed below.

Early-shear is the 'running-in' stage of wear at contacting asperities. Most solid rock runs exhibited slip-weakening with distinct associated drop of the wear and wear-rate (left fig.). The weakening distance before reaching steady-state, dW, is generally linearly proportional to the wear-drop distance, L0 (centre fig.). The work associated with this stage is fully explained by the energy dissipated in gouge surface area, thus the frictional-strength is dominated by wear resistance.

Steady-state is characterized by a continuous gouge layer, and the shear occurred in ‘three-body’ mode. Wear-rate drops by 1–3 orders of magnitude relatively to running-in (left fig.) with strong dependence on slip velocity. Wear is controlled by rock fracturing at the contact between the gouge layer and the solid rock, and the wear-rate is dominated by frictional-strength.

Runs with confined rock powders pose challenging results. First, all dry powder runs showed slip-strengthening, significant dilation, and velocity-strengthening at moderate velocities (v <0.2 m/s) (right fig.). Second, carbonate powders display slip-weakening, velocity-weakening and intense decomposition (CO2 emission) at high slip velocities (0.2 < V < 1.0 m/s). It appears that the frictional strength of confined powers is controlled shear distribution: localization leads to weakening whereas distributed shear (Riedel shears) leads to strengthening. The friction-wear relations is the subject of future work.


18

Day 1 — Session 3 — 14.25–14.50

SLOW EARTHQUAKES AND THE SPECTRUM OF FAULT SLIP BEHAVIORS

C Marone, J R Leeman, M M Scuderi, D M Saffer, C Collettini and P Johnson

Slow earthquakes and tectonic tremor represent an important conundrum in earthquake mechanics. Whereas standard earthquakes in Earth’s shallow crust are understood as catastrophic failure events that represent stick-slip frictional failure, the connection between frictional stability and slow earthquakes remains unclear. Moreover, unlike for regular earthquakes, where rupture speed is dictated by elastic wave speed of fault rocks, the factors that determine slip speed and rupture propagation velocity in slow earthquakes are unknown. Slow tectonic slip events and the related phenomena of low-frequency earthquakes represent a quasi-dynamic form of self-sustained, fault failure. Seismic and geodetic observations reveal a continuous spectrum of modes that unfold on timescales ranging from seconds to months but the mechanics of these phenomena are poorly understood. I discuss our recent laboratory observations for quartz fault gouge showing the full spectrum of slip modes including repetitive stick-slip with a wide range of slip velocities and event durations corresponding to standard and slow earthquakes. Our work shows that slow slip obeys the same frictional mechanics as regular earthquakes and that it occurs near the phase boundary predicted between stable and unstable failure, as controlled by the interplay of fault zone frictional properties and elastic stiffness. Our data show that slip velocity and event duration vary systematically as the stability threshold is approached, with 'normal' stick slip far from the transition, and increasingly slow slip transients, along with complex and chaotic modes of stick slip, near the threshold. This work suggest a general mechanism for slow and low frequency earthquakes that is consistent with theory, inferences about in situ fault conditions where slow earthquakes occur, and with the broad range of geologic environments in which these phenomena are observed.


19


THE BRITTLE-DUCTILE TRANSITION IN THE EARTH: A TRIP TO THE ZOO

Brian Evans1,2

1Massachusetts Institute of Technology, Cambridge, MA 02139, USA 2Department of Earth, Atmospheric, and Planetary Sciences

Studies of field structures, laboratory experiments, and mechanical analyses provide extensive evidence for a generalized transition in deformation style with increasing depth in the Earth. Depending on viewpoint, this transition can be described as a change in deformation mode, for example, as the tendency for localized cataclastic deformation in the upper portion of the Earth’s crust gives way to more distributed strain with increasing pressure and temperature. Alternatively, one may emphasize transitions in the dominant deformation mechanisms, e.g., cataclastic damage versus crystalline plasticity. Extensive recent studies of the cataclastic failure of porous rocks provide useful tools to understand the transition from distributed cataclasis to localized dilatant rupturing or to localized compaction banding. For the former transition, yield-cap models describe shear-enhanced compaction and include a transition from overall dilatancy to aggregate compaction. This model successfully rationalizes the phenomenology of inelastic deformation under a wide range of conditions, where several different physical mechanisms might dominate. Although the yield-cap formulation provides general insight into dependence of yield and hardening on the mean effective stress, there will be important differences in the constitutive behavior introduced by the identity of the deformation mechanisms, the appropriate internal damage variables, and the interactions amongst them. Natural inelastic deformation probably involves panoply of mechanisms, including cataclasis, stress corrosion, advective fluid transport, pressure solution, crystal plasticity, twinning, diffusion creep, boundary sliding, and chemical reactions. Combined with the boundary conditions of loading and kinematics, a rich variety of deformation structures could be expected in natural conditions. Thus, it is probably correct to think multiple transitions in mode, rather than a single general transition. With more detailed knowledge of mechanisms, these deformation modalities could be compared in multidimensional deformation maps, involving deformation rate, temperature, mean stress, pore-fluid conditions, and other internal variables, including grain size, pore geometry, porosity, and, of course, mineralogy. Current knowledge, although far from complete, does allow some tentative comparisons.


20

Day 1 — Session 4 — 15.40–16.05

THE IMPACT OF CEMENTATION ON MECHANICAL PROPERTIES AND PERMEABILITY OF LEITHAKALK CALCARENITE

P Baud1, U Exner2, T Reuschlé1 and M Lommatzsch3

1Institut de Physique du Globe de Strasbourg (UMR 7516 CNRS, Université de Strasbourg/EOST), 5 rue René Descartes F-67084 Strasbourg Cedex, France.

2Geological and Palaeontological Department, Natural History Museum, Burgring 7, 1010 Vienna, Austria.3Department of Geodynamics and Sedimentology, University of Vienna, Altanstrasse 14, 1090 Vienna, Austria

The analysis of deformation and failure in many sedimentary settings hinges upon a fundamental understanding of inelastic behavior and failure mode of porous carbonate rocks. However previous experimental studies on limestone showed that significant variability in the mechanical behavior can be observed in this rock type, even for samples coming from the same sedimentary unit. While differences in the degree cementation are clearly one of the parameters at the origin of this variability, the impact of cementation on the brittle-ductile transition in carbonate rocks has to our knowledge never been systematically studied.

In this work, we tried to fill the gap and performed 30 uniaxial and conventional triaxial experiments on samples of Leitha calcarenite, a Miocene carbonate grainstone from the Vienna Basin, Austria. We selected samples in one quarry with different degrees of cementation and porosities ranging from 18 to 31%. They were water-saturated and deformed in drained conditions, at room temperature, constant strain rate and at effective pressures ranging from 5 to 200 MPa. Brittle behaviour was only observed in uniaxial conditions and in experiments performed at very low effective pressures. Most of our triaxially deformed samples failed by shear-enhanced compaction. Our new data reveal a systematic decrease of both the Uniaxial Compressive Strength and the onset inelastic compaction with porosity (less cementation) and also suggest that both quantities could be predicted by a very simple empirical expression for Leitha calcarenite. SEM observations as well as high resolution X-ray Tomography (µCT) data suggested that the differences in strength was not so much related to the variation in macroporosity, but rather a difference in intermediate and small sized pores within the individual bioclasts. In the studied porosity range, pore-emanated microcraking is the main micromechanism leading to brittle failure, while cataclastic pore-collapse and grain crushing are the main mechanisms of inelastic compaction, in qualitative agreement with existing theoretical models.

The permeability of intact Leitha calcarenite was measured using gas over the whole range of porosities, and compared to a new series of permeability measurements on a wide selection of carbonates. Our new permeability data showed that for a comparable porosity, the permeability of Leitha calcarenite is significantly larger than that of other carbonates, and also that the permeability of this rock, unlike its mechanical strength, is not significantly influenced by the degree of cementation. This observation can be explained by the preservation of macroporosity by grain-coating cement, while especially intra-grain micropores are sealed by progressive cementation, in agreement with our microstructural observations and µCT data.


21

Day 1 — Session 4 — 16.05–16.30

DEM SIMULATIONS OF COMPACTION BANDS IN SANDSTONES, AS RELATED TO DIFFERENCES BETWEEN LAB AND FIELD OBSERVATIONS

G Marketos1, M P A van den Ende and A R Niemeijer

1Department of Geoscience, Utrecht University, The Netherlands

Compaction bands are zones of localised compaction that accommodate little or no shear and have been observed in porous rock both in laboratory experiments (e.g. [1]) and in the field (e.g. [2]). As they can form in sandstone, often the source rock for hydrocarbons, understanding the conditions under which they form can be very important. For example they can significantly reduce permeability of the reservoir rock.

This contribution is motivated by the fact that there seems to be a difference in compaction bands observed in the field and in the laboratory. The micromechanism by which compaction is accommodated in the band is through grain breakage in the laboratory, while more gentle grain deformation seems to have been observed in the field. Here we present results from Discrete Element Method (DEM) simulations that investigate these two different micro-scale mechanisms that can lead to compaction banding. Grain crushing compaction bands have been successfully reproduced in DEM ([3]) while ones due to pressure solution remain more elusive. Similarities between the two situations will be highlighted, and reasons why one mechanism might prevail in the laboratory or the field will be discussed. Micro-conditions necessary to reproduce pressure-solution compaction bands in the simulations will be discussed with the aim of better understanding their occurrence. This could eventually lead to prediction of their existence inside sandstone formations.

References[1]. Mollema, P N and Antonellini, M A. 1996. Compaction bands: a structural analog for anti-mode I cracks in Aeolian sandstone. Tectonophysics 267, 209–228.

[2]. Baud, P, Klein, E and Wong, T-F. 2004. Compaction localization in porous sandstones: spatial evolution of damage and acoustic emission activity. Journal of Structural Geology 26, 603–624.

[3]. Marketos, G and Bolton M D. 2009. Compaction bands simulated in Discrete Element Models. Journal of Structural Geology 31, 479–490.


22

Day 1 — Session 5 — 16.55–17.20

MECHANISM AND MODELING OF BRITTLE-DUCTILE FAILURE OF POROUS ROCKS

E Shalev1, W Zhu2 and V Lyakhovsky1

1Geological Survey of Israel, 30 Malkhe Israel Street, Jerusalem 95501, Israel2Department of Geology, University of Maryland, 237 Regents Drive, College Park, MD 20742, USA

Porous rocks fail by shear localization or by cataclastic flow depending on loading conditions. We quantify the role of different deformational mechanisms in a model which accounts for the coupling between damage accumulation, dilation, and compaction during loading of sandstones. Formulation includes three different deformation regimes: I) elastic deformation characterized by material strengthening and compaction; II) cataclastic flow characterized by damage increase and compaction, and III) brittle failure characterized by damage increase, dilation, and shear localization. Using a three dimensional numerical model, we simulate the deformation behaviour of cylindrical porous Berea sandstone samples under different confining pressures. The obtained stress, strain, porosity changes as well as macroscopic deformation features well reproduce the laboratory results. The model predicts different rock behaviour as a function of confining pressures. The elastic and brittle regimes associated with formation of shear and/or dilatant bands occur at low effective pressures. The model successfully reproduces cataclastic flow and homogeneous compaction under high pressures. Complex behaviour with overlap of common features of all regimes is simulated under intermediate pressures, resulting with localized compaction or shear enhanced compaction bands. Numerical results elucidate three steps in the formation of compaction bands: 1) dilation and subsequent compaction of elements on a localized plane, 2) formation of high angle shear enhanced compaction band, and 3) formation of horizontal pure compaction band.


23

Day 1 — Session 1 — 17.20–17.45

MICROPOROSITY DISTRIBUTION AND BRITTLE TO DUCTILE TRANSITION IN OOLITHIC CARBONATE ROCKS

J B Regnet1, C. David2, J Fortin1, P Robion2, Y Makhloufi3 and P Y Collin3

1Laboratoire de Géologie de l’Ecole Normale Supérieure – PSL Research University - UMR CNRS 8538, Paris, France 2Université de Cergy-Pontoise, Laboratoire Géosciences et Environnement Cergy, France

3Université de Bourgogne, UMR CNRS 6282 Biogéosciences, Dijon, France

The mechanical behavior of oolithic carbonate rocks was investigated for selected rocks with two different microstructural attributes: uniform (UP) and rimmed (RP) distribution of microporosity within ooids. These oolithic carbonate rocks are from the Oolithe Blanche formation, a deep saline aquifer in the Paris Basin, and a possible target for CO2 sequestration and geothermal production. Samples of similar physical properties (porosity, grain diameter, cement content) but different microporosity textures were deformed under triaxial configuration, in water saturated conditions, at 28 MPa of confining pressure, 5 MPa of pore pressure and at a temperature of 55°C. During the experiments, acoustic velocities were monitored, and permeability was measured. The results show that the mechanical behavior of these microporous carbonates are strongly controlled by the microporosity distribution within the grains, at the origin of variations in elastic properties, mechanical strength and failure mode. The lower velocities measured in UP samples indicate a larger compliance of the whole structure. The mechanical response indicates that UP samples are characterized by a ductile behavior whereas RP samples display a brittle behavior. Using a conceptual model for the failure envelope of both rocks, our observations can be accounted for if one considers a significant variation of the critical pressure P*, with UP samples having a lower P* than RP samples. The permeability evolution under stress was interpreted using a revised Kozeny Carman equation, showing that fluid flow is strongly affected by the tortuosity of the pore space, which is controlled by the microporosity distribution within the ooids. This study brings new insight into the parameters controlling the physical and mechanical response of oolithic carbonates, and the possible impact on production of geothermal energy at depth or storativity for CO2 sequestration operations.


24

Day 1 — Session 5 — 17.45–18.10

CONTRASTING FAILURE MODES OF FOLDED GNEISS REVEALED BY MECHANICAL, MICROSEISMIC AND MICROSTRUCTURAL DATA

F Agliardi1, S Vinciguerra2,3, M Dobbs3 and S Zanchetta1

1University of Milano-Bicocca, Earth and Environmental Sciences, Piazza della Scienza 4, Milano, 20126, Italy

2Department of Geology, University of Leicester, University Road, Leicester, LE1 7RH, UK

3British Geological Survey, Environmental Science Centre, Keyworth, Nottingham, NG12 5GG, UK

It is well established that planar rock fabric anisotropy is a major control on the mechanical behaviour of rocks. It is known to influence brittle rock strength and deformability depending on its orientation with respect to loading directions and on the presence of weak phases such as phyllosilicates (Shea and Kronenberg, 1993). Furthermore, a recent experimental study in uniaxial compression (Agliardi et al., 2014) suggested that inherited folded anisotropic fabric affects rock failure behaviour in a completely different fashion, due to the interplay of the folded foliation and an 'axial plane anisotropy'. However, this has been shown in unconfined compression conditions, and there is a need to investigate whether the effects persist at increasing depth. Thus we carried out triaxial compression laboratory experiments, monitoring the microseismic output (AE) and undertaking detailed micro-structural analyses of the resulting failure patterns. We tested Monte Canale gneiss from Val Malenco (Central Italian Alps), which is characterized by relatively low phyllosilicate content and foliation that is folded on a mm to the cm-scale. We performed experiments using a servo-controlled hydraulic load frame equipped with triaxial pressure vessel (MTS 815 rock mechanics testing system) on 25 cylindrical specimens (D: 54 mm; H: 110–130 mm). Each specimen was first analysed in terms of density, porosity, fold geometry and orientation of foliation and fold axial planes. Specimens were instrumented with strain gauges and deformed at confining pressures of 40 MPa (19 specimens) and 120 MPa (6 specimens), and constant axial strain rate of 5*10-6 s-1. We measured P- and S-wave velocities and AE using piezoelectric transceivers covering the 100–1200 kHz range. We performed microscale analyses of resulting fracture process zones combining optical microscopy, X-ray MicroCT and SEM-BSE on resin-impregnated sub-samples. Specimens failed in three distinct modes characterized by different combinations of neat shear planes parallel to foliation, fractures parallel to fold axial planes, or less localized mm-scale brittle shear zones at both confinements. The different failure modes are mirrored by distinct stress-strain relationships, microcracking distribution and acoustic emission features (overall activity, rates, frequency/amplitude patterns). Failure modes involving the quartz-dominated axial plane anisotropy are characterised by i) high values of peak strength and axial strain; ii) wide fracture/damage zones and high microcrack density, mirrored by iii) intense and continuously increasing AE activity compared to failure modes where fracture process zones are localized along the mica-dominated foliation. Our experimental and microstructural observations confirm and extend previous findings. They support the evidence that, even at increasing depths, the key control on deformation and failure processes is the folded microfabric, by the activation of competing Q-dominated and M-dominated crack nucleation/propagation mechanisms eventually resulting in very contrasting mechanical behaviours of the same rock type.


25

TALKSDAY 2


26


LABORATORY EARTHQUAKES

R Paul Young

Lassonde Institute, University of Toronto, Canada

Since the early pioneering work of Brace et al. (1963) and others, laboratory studies have been used to provide a well-controlled environment for studying fracture processes in rocks. Numerous investigators have shown how laboratory studies can yield crucial information in the understanding of earthquake processes. Sophisticated Acoustic Emission (AE) monitoring has also been used to map out the details of rock failure and unstable slip on existing fractures (e.g. Lockner et al. 1991). More recently, quantitative seismology techniques have been used with calibrated AE sensors by McLaskey et al (2013) at the USGS and by Goodfellow and Young (2014) to demonstrate that the scaling process for earthquake physics extends down at least to the grain scale.

Governments and the public are increasingly concerned over induced and triggered seismicity associated with oil and gas recovery, hydraulic fracturing and also potential ground water contamination associated with these activities. Deep geological disposal of radioactive waste is also an issue that highlights the need to characterize/quantify induced fractures and rock permeability during the excavation and sealing of repositories. This presentation will discuss these issues and show how laboratory earthquakes at various scales can provide a method for geophysical characterization of induced fracturing, seismicity and fluid flow in fractured rock and contribute to understanding how these processes scale from laboratory to in situ conditions.


27

Day 2 — Session 6 — 09.50–10.15

MECHANICAL AND PETROPHYSICAL STUDY OF FRACTURED SHALE MATERIALS

Audrey Bonnelye1,2, Alexandre Schubnel2, Christian David1, Pierre Henry3, Yves Guglielmi3, Claude Gout4 and Pierre Dick5.

1Université de Cergy Pontoise2Ecole Normale Supérieure de Paris

3CEREGE, Marseille4TOTAL5IRSN

Studying the mechanical and physical properties of shales is of major importance for the understanding of cap-rock reservoir, nuclear waste disposal or upper crustal fault hydro-mechanical behavior. In particular, relationships between applied stress, textural anisotropy and transport properties are critical both in intact and fractured shales. These relations can be investigated in the laboratory in order to have a better understanding on in-situ mechanisms: here we used intact and naturally fractured samples from the Toarcian shale formation of the Tournemire Underground Research Laboratory (URL), southern France Conventional triaxial tests were performed in order to determine the elasto-plastic yield envelope on three sets of samples with different orientations relative to bedding (0°, 45°, and 90°). For each set, six experiments were carried out at increasing confining pressures (2.5, 5, 10, 20, 40, 80MPa). Experiments were performed under nominally drained conditions, at strain rates ranging between 5x10-7 s-1- 1x10-5 s-1 up to failure. During each experiment, P and S wave elastic velocities were continuously measured along different directions, in order to monitor the evolution of elastic anisotropy in our material. Our results show that the orientation of principal stress components relative to the bedding plane plays an important role on the brittle strength. Minimum strength is observed for samples when maximum compressive stress is oriented 45° to bedding. The relative weakness of samples oriented at 45° increases with confining pressure, but shows no strain rate dependence. On the contrary, samples oriented at 0° and 90° exhibit large strain rate dependence. We interpret this result as the cohesive strength (and fracture toughness) being strain rate dependent, while the friction coefficient of Tournemire shale being strain rate independent over the range of velocities investigated here. We also observed that brittle failure is preceded by the development of P wave anisotropy, due to both crack growth and mineral re-orientation. P wave anisotropy development is largest for samples with s1 oriented perpendicular to bedding, with anisotropy being largest at the onset of rupture. Anisotropy reversal can be observed at the highest confining pressures. For samples with s1 oriented parallel to bedding, the P wave anisotropy development is weaker, associated with shorter axial strain (and plastic deformation) before and smaller stress drops at rupture. For both of these orientations, Thomsen parameters ( and ) were inverted from the elastic wave data in order to quantify the evolution of elastic anisotropy during deformation.

We also studied the evolution of P-waves velocities of samples from a borehole crossing a natural fault as a function of the distance to the fault core position. We observed an evolution of P wave velocity anisotropy and found similarities with our triaxial experiments with the re-orientation of layers and coalescence of several fractures to form the fault core.


28

Day 2 — Session 6 — 10.15–10.40

ANISOTROPY AND FRACTURE PROPAGATION IN SHALE WITH ELEVATED CONFINING PRESSURES

M Chandler2,1, P Meredith1, N Brantut1, B Crawford3, J Mecklenburgh2 and E Rutter2

1Rock and Ice Physics Laboratory, Department of Earth Sciences, University College London, UK.2School of Earth, Atmospheric and Environmental Sciences, The University of Manchester, UK.

3ExxonMobil URC, 3120 Buffalo Speedway, Houston, TX 77098

The use of hydraulic fracturing to recover shale-gas has drawn attention to the fundamental fracture properties of gas-bearing shales. Fracture propagation trajectories in these materials depend on the interaction between the anisotropic mechanical properties of the shale and the anisotropic in-situ stress field. However, there is a general paucity of available experimental data on their anisotropic mechanical and physical properties.

The fracture toughness of a linear elastic material is defined as the critical value of the stress intensity factor; KIc, beyond which rapid, catastrophic crack growth occurs. However, shales deviate significantly from linear elasticity, and exhibit marked hysteresis during cyclic loading. Helpfully, this hysteresis enables the calculation of a ductility coefficient using the residual displacement after each successive loading cycle. A ductility-corrected fracture toughness value, KIc

c, was calculated using this coefficient following the ISRM suggested method. In Mancos shale this ductility correction can be as large as 60%.

Here we report experimental measurements of the tensile strength () and for Mancos shale and Whitby mudstone determined at ambient conditions using the Brazil Disk and Short-Rod methods respectively. Measurements were made in all three of the principal fracture orientations; Arrester, Divider and Short-Transverse. Significant anisotropy is observed in both and KIc

c measurements, with KIcc in the Divider and Arrester orientations being around 1.8

times that in the Short-Transverse orientation in Mancos shale. For both and KIcc, the Short-Transverse orientation,

where the fracture propagates parallel to the bedding, is found to have significantly lower values than the other two orientations. In Mancos shale, a bimodal distribution in KIc

c, is observed in this orientation, with two distinct values recorded repeatedly. This has been interpreted as being due to the layered nature of the shale material.

We also report a method for performing Short-Rod effective KIcc, experiments at elevated (jacketed) confining

pressure, and measurements made on Mancos shale and comparison materials, with KIcc, increasing approximately

four-fold between ambient conditions and experiments conducted at 30MPa. KIcc, anisotropy was observed to

decrease substantially with increasing confining pressure over the same range, suggesting that the material behaves near-isotropically at depth.

In order to investigate the geometry of propagating hydraulic fractures in shale, fluid injection experiments can be used involving an injection fluid visible in x-ray tomographs so that fracture geometry can be observed after the experiment is complete.


29

Day 2 — Session 6 — 10.40–11.05

THE ROLE OF PHYLLOSILICATE-RICH MYLONITIC FABRIC ON DEFORMATION, FAILURE MODE AND FAULT WEAKNESS: NEW INSIGHTS FROM ROCK

DEFORMATION LABORATORY EXPERIMENTS

F Bolognesi1, S Vinciguerra2,3, A Bistacchi1 and M Dobbs3

1Department of Earth and Environmental Sciences, Bicocca University of Milan,Italy 2Department of Earth Sciences, University of Turin, Italy.

3British Geological Survey, Environmental Science Centre, Keyworth, Nottingham, UK

One mechanism explaining the nucleation and propagation of weak faults with non-Andersonian attitude is the mechanical anisotropy of phyllosilicate-rich mylonitic rocks. We characterized the mechanical anisotropy and (micro-)failure modes of phyllosilicate-rich (30%) mylonites from the Grandes Rousses Massif (Helvetic-Dauphinois Domain, French Alps), deformed under brittle conditions after exhumation from metamorphic conditions. We performed uniaxial (UCS) and triaxial (TXT) tests varying the 1/schistosity angle at confining pressures of 60 and 120 MPa. Tests were carried out at a constant strain rate of 6*10-6 s-1 and microseismicity in terms of acoustic emissions (AE) are measured during the rock deformation experiments

UCS at 90° show high strength and failure mode characterized by both low-angle segments along schistosity and high-angle ones cutting quartz-feldspar layers. UCS at 0° show lower strength and axial splitting along schistosity.

TXT cover a complete range of inclinations, from 0° to 90°, at confining pressures of 60 MPa and 120 MPa. Maximum strength is achieved at 0°, and minimum strength, attained at 45°, is around 50% with a significant strength anisotropy. The 'fault zone' develops mainly along schistosity for tests at 20–70°, whilst Andersonian shear fractures are observed at 0–20° and 70–90°.

Microstructural observations further suggest that micaschists foliation may have a primary role in the enucleation, propagation and coalescence of microcracking during dilatant failure. Foliated rocks are found in a wide range of 1/schistosity angles and, because the failure mode is controlled by the slip along the schistosity, a significant deviation from the prediction of Anderson’s theory is observed depending upon the angle to 1 at which new macroscopic fault zones enucleate. AE mirror the mechanical behaviour with higher rate occurrence related to higher mechanical damage.


30

Day 2 — Session 7 — 11.30–11.55

MODELLING OF THERMO-HYDRO-MECHANICAL COUPLED PROCESSES FOR FLUID-BEARING RESERVOIRS

A B Jacquey1,2, M Cacace1, G Blöcher1 and M Scheck-Wenderoth1,2

1GFZ, German Research Centre for Geosciences, Potsdam, Germany2RWTH Aachen University, Aachen, Germany

Understanding the coupling between thermo-hydro-mechanical processes in saturated porous media is of interest for several geo-energy related studies such as geothermal power production, energy storage or enhanced petroleum and gas recovery. Indeed, characterisation of pore pressure, temperature and stress distributions within a reservoir during injection and production of fluid can help to understand reservoir behaviour and constrain reservoir productivity. This study presents an approach for describing the coupling between thermo-hydro-mechanical processes within fluid-bearing reservoirs. This approach includes, among diverse processes, evolution of transport properties (porosity and permeability) as induced by pore pressure and temperature changes formulations consistent with the theories of poro- and thermoelasticity.

The above formulations have been implemented into a Finite Element Method based software to carry on forward thermos-hydro-mechanical simulations. Using calibration parameters as provided by laboratory tests provided under drained hydrostatic compressions, the consistency of these formulations is first assured and verified. Once calibrated against available data, the approach is applied to real case geothermal reservoirs of complex geometries, comprising fault zones, fractures and heterogeneous geology as shown in Fig. 1. Forward simulations are then used to make quantitative predictions on the behaviour of the reservoir in terms of productivity and sustainability.

Fig. 1 Porosity distribution induced by injection of cold geothermal fluid in one geological layer of the Groß Schönebeck geothermal

reservoir (40 km north of Berlin, Germany).


31

Day 2 — Session 7 — 11.55–12.20

IN SITU STATE OF STRESS AT IODP HOLE C0002A NEAR NANKAI TROUGH DETERMINED BY A STOCHASTIC PROCESS OF SONIC VELOCITY AND

BREAKOUT WIDTH LOGGING DATA

I Song1, C Chang2 and H Lee1

1Deep GeoEnvironment Research Centre, KIGAM, Daejeon 305–350, Republic of Korea2Dept of Geology, Chungnam National University, Daejeon 305–764, Republic of Korea

The Nankai Trough Seismogenic Zone Experiment (NanTroSEIZE) is a long-term, multi-stage scientific drilling project launched for investigating fault mechanics and seismogenesis along subduction megathrusts along the Nankai accretionary prism, SW Japan. One main key to the plate boundary mechanics is understanding the absolute mechanical strength and the in situ state of stress along the thrust zone. The stress regime in the shallow portion of the Nankai accretionary prism has been well documented by the analysis of borehole failure detected in several vertical boreholes. During Integrated Ocean Drilling Program (IODP) Expeditions 314/315/316 and 319 conducted as part of NanTroSEIZE, the orientation of in situ stress was revealed from drilling-induced compressive failures (breakouts) and tensile failures (DITFs) observed in borehole wall resistivity images. The orientation of the maximum horizontal stress (Shmax) is subparallel to the plate convergence direction in the outer wedge and also in the centre of the Kumano basin. However, the Shmax direction is rotated ~90 degrees near the margin of Kumano basin at Site C0002. Borehole wall images and sonic logging data obtained from the riserless-drilling of C0002A show a wide distribution of breakout width and sonic velocity even at a short interval of depth. The small-scale but frequent variation in breakout width in a short section of borehole wall is due to heterogeneous rock strength rather than a correspondingly frequent change in far-field stress. In this paper we consider the probability distribution of rock strengths and breakout widths in a given section of wellbore, which is large enough to analyze the logging data in a statistical manner but small enough to make sure that the far-field stresses are to be uniform, in order to determine the magnitudes of in situ stresses. Assuming the normal distribution of uniaxial compressive strength (UCS), which is estimated empirically from sonic velocity logs (Chang et al., 2006, JPSE; 2010, G3), we calculated the probability distribution of breakout width for given sets of the maximum and the minimum horizontal principal stresses (SHmax and Shmin, respectively) for every 30 m depth interval. The same procedure was repeated for various combinations of the two horizontal principal stress magnitudes. Then the objective function with two variables, SHmax and Shmin, was obtained from the total misfits between the observed and the calculated occurrence distributions of breakout width. Finally we were able to determine the best solution of SHmax and Shmin with the minimum total misfit. The results from this new approach of stress estimation are comparable with previous other results (e.g., Chang et al., 2010, G3; Lee et al., 2013, MPG) which gave a wide range of in situ stress magnitudes. According to the new stochastic prediction of in situ stress, the Kumano Basin is in a subcritical state of normal faulting as a series of normal faults near the wellbore are shown in the seismic profile. Underneath the unconformity, the normal fault stress regime sharply changes to the strike-slip stress regime in the old accretionary prism by rapidly increasing in both horizontal stresses. This stochastic model is prominent because it gives not only both values of SHmax and Shmin simultaneously but also information about statistical reliability of the determined values quantified by sensitivity and uncertainty. Our result shows that the two stress magnitudes in Nankai accretionary prism are not completely independent in terms of sensitivity, suggesting that other independent measure of one of the two stresses might be definitely useful (e.g., from leak-off test).


32


33

TALKSDAY 3


34


MODELS AND MEASUREMENTS: CASE STUDIES IN THE VISCOELASTIC AND POROELASTIC BEHAVIOUR OF ROCKS

Ian Jackson

Research School of Earth Sciences, Australian National University, Canberra ACT Australia

The posing of hypotheses and their testing against observations is the essence of the scientific method. Elegant theory has been developed to describe the role of grain-boundary sliding in the high-temperature viscoelastic behaviour of polycrystalline materials. Likewise, the poroelastic relaxation of fluid-saturated media has attracted sustained theoretical interest. In each of these fields, however, the testing and refinement of such theoretical models for a robust understanding of material behaviour is dependent upon the ongoing development and application of complementary experimental methods that allow broadband observation of the inevitably frequency-dependent mechanical behaviour. In particular, forced oscillation methods are now providing critical constraints at seismic frequencies and microstrain amplitudes. The first case study is the high-temperature viscoelastic behaviour of polycrystalline olivine in which departures from ideal elasticity, and the resulting frequency dependent behaviour, are associated with the stress-induced migration of crystal defects including point defects, dislocations and grain boundaries, or the stress-induced redistribution of an intergranular melt phase. Representative results for natural peridotites and synthetic analogues will be presented — highlighting aspects of the observed variations of shear modulus and dissipation with frequency, temperature, grain-size, dislocation density, melt fraction, and the concentration of water. Implications for seismic wave dispersion and attenuation in the Earth’s upper mantle will be addressed. In the second case study, of poroelastic relaxation, theoretical models predict frequency dependent wave speeds and attenuation associated with the stress-induced redistribution of the pore fluid. Results will be reported from an ongoing study of synthetic glass specimens with contrasting crack-pore microstructures. Such materials have been tested with complementary methods which in combination span more than six orders of magnitude in frequency: ultrasonic wave propagation (MHz), resonant bar methods (kHz), and forced oscillation (mHz-Hz). The new results are interpreted as variously probing the saturated-isolated, saturated isobaric and drained regimes of poroelastic behaviour. In each case study, the necessarily iterative process of refining theoretical models to meet observational constraints will be highlighted.


35

Day 3 — Session 8 — 09.50–10.15

AN APPARATUS TO MEASURE FREQUENCY-DEPENDENT POISSON RATIO OF FLUID-SATURATED SANDSTONES

L Pimienta1, J Fortin1 and Y Guéguen1

1laboratoire de Géologie de l’ENS — PSL Research University — UMR8538 du CNRS.

Although seismic wave dispersion and attenuation are known to occur in sedimentary rocks, measuring this effect experimentally remains challenging. A new experimental apparatus and protocol has been calibrated at the Laboratoire de Géologie of ENS in order to measure both Young modulus and attenuation as a function of frequency in rock samples. Yet, when investigating the elastic properties of isotropic samples, another elastic constant is needed. The apparatus is thus tested for the measurement of Poisson ratio of fluid-saturated rock samples.

Three standard samples (i.e. Glass, Gypsum and Plexiglas) are used to test the dependence to pressure and frequency of the experimental set-up. The samples are chosen because, in the conditions P-T of investigation, (i) all three samples are not pressure-dependent; (ii) Glass and Gypsum are not frequency-dependent; and (iii) Plexiglas is a viscoelastic material that has a characteristic frequency-dependent behaviour.

Figure 1: Dependence to (a) pressure and (b) frequency of the Poisson ratio of Glass, Gypsum and Plexiglas.

Poisson ratio of the three samples is measured as a function of pressure (Fig.1a) and frequency (Fig.1b). The measurements are compared with the values from the literature. For all samples, Poisson ratio is not dependent on pressure, and no dependence to the measuring frequency is observed for Glass and Gypsum samples (Fig.1a). Reversely, the Plexiglas sample’s Poisson ratio decreases as frequency increase, from 0.40 at lowest frequency down to about 0.36 at highest frequency (Fig.1b).

The calibrated approach is used to measure the Poisson ratio of a Fontainebleau sandstone sample of 7% porosity and 10–14 m2 permeability is measured under dry, water- and glycerine-saturated conditions. Interestingly, while very small dependence to pressure and frequency is measured under dry conditions, this sample’s Poisson ratio shows very large dependences when saturated by either water or glycerine. For all effective pressures, frequency-dependent bell-shaped variations are measured. Using fluid flow and effective medium theories, it will be shown that the measured variations are consistent with the two transitions from drained to undrained and from undrained to unrelaxed regimes.


36

Day 3 — Session 8 — 10.15–10.40

STATIC AND DYNAMIC POROELASTIC MODULI OF MALM CARBONATE

Alireza Hassanzadegan1, Romain Guérizec2, Thomas Reinsch1, Guido Blöcher1 and Günter Zimmermann1

1Basin Analysis, GFZ German Research Centre for Geosciences, Helmholtz Centre Potsdam2Géosciences, Université Montpellier II

Injection and production of water in geothermal reservoirs leads to pore pressure and temperature changes, resulting in changes in the stress acting on reservoir and surrounding rocks. The geomechanical behavior of geothermal reservoir rocks, e.g. in situ stress changes and deformation, can be characterizes if mechanical properties of the rocks are known. This study provides poroelastic properties (e.g., drained bulk modulus, Biot coefficient) of the Malm carbonate, the target reservoir formation within foreland Molasse basin, south west of Germany.

Rock samples were collected from outcrops at the surface, representing an analogue to the geothermal reservoir rock. The hydraulic and poroelastic properties were measured. The porosity was measured, using three different methods; Helium pycnometer, mercury porosity and weighting methods. Drained jacketed experiments were performed at 30 and 60°C, and unjacketed experiments at 60, 90 and 120°C. The static and dynamic elastic moduli of carbonate rocks were measured using stress-strain measurements and acoustic velocity measurements, respectively. The temperature was increased stepwise from 30 to 120°C, and measurements were performed at isothermal conditions. The confining pressure and pore fluid pressure were controlled independently to simulate the in situ stress conditions of a buried rock. The pore fluid pressure was kept constant at 2 MPa and confining pressure was cycled between 3 to 80 MPa.

The porosity of rock decreased from 10.85% at 2 MPa to 10.68 % at 80 MPa with increasing effective pressure. The static drained bulk modulus was increasing from 20.8 GPa to 48.2 GPa with increasing effective pressure, the dynamic drained bulk modulus ranged between 40 to 42.5 GPa. The results are useful to correlate the static and dynamic moduli and to populate the geomechanical models with static elastic moduli.


37

Day 3 — Session 8 — 10.40–11.05

RECOVERY OF DAMAGE AND ELASTIC WAVE SPEEDS IN DEFORMED LIMESTONE

Nicolas Brantut

Department of Earth Sciences, University College London, UK

Limestone samples were deformed up to 5% inelastic axial strain at an effective confining pressure Peff= 50 MPa, in the cataclastic flow regime, and subsequently maintained under constant static stress conditions for extended periods of time while elastic wave speeds and permeability were continously monitored. During deformation, both seismic wave speeds and permeability decrease with increasing strain, due to the growth of sub-vertical microcracks and inelastic porosity reduction. During the static hold period under water-satured conditions, the seismic wave speeds recovered gradually, typically by around 5% (relative to their initial value) after two days, while permeability remained constant. The recovery in wave speed increases with increasing confining pressure, but decreases with increasing applied differential stress. The recovery is markedly lower when the samples are saturated with an inert fluid as opposed to water. The evolution in wave speed is interpreted quantitatively in terms of microcrack density, which shows that the post-deformation recovery is associated with an decrease in effective microcrack length, typically of the order to 10% after two days. The proposed mechanism for the observed damage recovery is microcrack closure due to a combination of backsliding on wing cracks driven by time-dependent friction and closure due to pressure-solution at contacts between propping particles or asperities and microcrack walls. The recovery rates observed in the experiments, and the proposed underlying mechanisms, are compatible with seismological observations of seismic wave speed recovery along faults following earthquakes.


38

Day 3 — Session 9 — 11.30–11.55

EVIDENCE FOR PERMEABILITY HETEROGENEITY IN A VOLCANIC CONDUIT

J Farquharson1, M J Heap1, Y Lavallée2, N R Varley3 and P Baud1

1Laboratoire de Déformation des Roches, Équipe de Géophysique Expérimentale, Institut de Physique de Globe de Strasbourg (UMR 7516 CNRS, Université de Strasbourg/EOST), 5 rue René Descartes, 67084 Strasbourg cedex, France.

2Earth, Ocean and Ecological Sciences, University of Liverpool, Liverpool L69 3GP, United Kingdom.3Facultad de Ciencias, Universidad de Colima, Mexico.

Volcanic conduits, domes, and edifices are by nature highly heterogeneous. During a recent field campaign at Volcán de Colima (Mexico), a range of rocks of explosive and extrusive origin were collected from sites on the volcano. The collected rocks all contained textural or structural heterogeneities, typically in the form of a planar or subplanar band of notably different colour, ranging from mm-scale thicknesses to zones over 5 cm thick. Microstructural and chemical analyses allowed us to separate the collected samples into three categories based on the inferred genesis of the feature of interest: banded pumice; shear strain-induced dilation or densification; and variably sintered fractures.

In the first case, alternating higher and lower porosity bands in pumice are inferred to be preserved evidence of inhomogeneous bubble expansion during the decompression of ascending magma, based on microstructural bubble textures and evidence of crystal breakage.

The second category consists of flow-banded blocks, indicating differentials in shear strain within the conduit over very small distances (i.e. decimetre scale and smaller), for example due to inhomogeneous magma flow in the conduit. Where the original material (i.e. the magma) contained very little porosity (<10%) then this strain localisation induces dilatancy, in that the bands are more porous than the surrounding material (up to around 20%). However, in an initially more porous sample (around 15% porosity), a local increase in shear strain appears to reduce porosity, leaving dense flow bands containing few, deformed bubbles.

The final category includes samples which have been subject to local viscous shear strains and/or strain rates sufficient to cause magma fragmentation. This process has been followed by viscous sintering of the ash-sized fracture plane material. Variable sintering efficiency and presence of cristobalite implies different depths and/or timescales of fracture and healing processes.

Permeability was measured using a steady-state gas permeameter on host rock cores and samples cored so that the feature runs parallel or perpendicular to the long axis. These data were used to estimate band permeability w.r.t. the host rock. Bands in pumice were not found to significantly influence permeability, despite containing porosities much lower than the host material (~20–30% compared to ~50–60%). High initial permeability (~10–13 m2) assures that effective fluid pathways are already established. Localised strain (without fracture) can significantly increase permeability (up to 3 orders of magnitude) when the initial host porosity is low ( 10%). Where the starting porosity is higher ( 10%) the bands are of similar or lower porosity than the surrounding medium. In the latter case, shear strain-induced densification of magma creates a barrier to fluid flow, the permeability of the dense band approximately an order of magnitude lower than the host rock. In the third category, the maturation of the sintered bands dictates the effect of the band on bulk porosity. Generally these variably-healed fractures have permeabilities in the range of 10–13 m2 (compared to the host media in the range of 10–16 m2), indicating that they comprise highly effective pathways for fluid flow.

Where fluid pathways (such as fractures in magma) are formed, they can provide an outgassing route for magmatic volatiles from a volcanic system. However, strain-induced barriers to flow (viscous densification) could have the opposite effect: trapping exsolved volatiles, allowing the build-up of pressure. Localised permeability heterogeneities likely constitute important controls on pressure evolution in shallow magmatic systems and thus, on the dynamic of volcanic eruptions.


39

Day 3 — Session 9 — 11.55–12.20

TENSILE STRENGTH CONSTRAINTS ON THE TRIGGERING OF VULCANIAN EXPLOSIONS AT SANTIAGUITO VOLCANO, GUATEMALA

A J Hornby1, Y Lavallée1, J E Kendrick1, A S D Collinson2 and J Neuberg2

1School of Earth, Ocean and Environmental Sciences, University of Liverpool, L69 3GP, UK2School of Earth and Environment, The University of Leeds, Leeds UK

Gas- and ash explosions at Santiaguito volcano occur at regular intervals, exiting through arcuate fractures in the summit dome of the Caliente vent. Infrasound, tomography, ground deformation and seismic monitoring have constrained the source for these explosions at 250–300 m depth below the dome. The explosions liberate a pressurized, gas-rich domain below a denser, impermeable magma plug. Rapid decompression of this gas-rich ‘pocket’ occurs through fracturing of the overlying plug, causing connection with atmospheric pressure at the dome surface, leading to an explosion. We surmise that the dominant fracture mode at these shallow depths is tensile due to the volumetric strain exerted by a pressurizing source, however a component of shear is also detected during explosive events. Fractures may propagate downwards from the dome surface (due to greater magma stiffness and lower pressure) or upwards from the gas pocket due to higher strain rates at the deformation source in the case of viscous deformation. In order to constrain the origin of fractures leading to explosions we conducted high-temperature Brazilian tensile stress tests on rocks from the Caliente dome at Santiaguito volcanic complex over a range of strain rates. At experimental temperatures of 750–800°C the tested dome rocks show a transition from the brittle fracturing to viscous deformation. In this regime, the material becomes highly sensitive to strain rate, showing a full range of response from fully elastic to viscous behaviour. The total strain before fracture initiation increases as viscous deformation becomes more dominant (at lower strain rates and higher temperatures). This leads to constraints on timescales for fracture propagation for a given temperature-strain rate scenario. By combining rock mechanical properties and monitored cycles of ground deformation using a numerical model we can estimate the stress conditions and strain rates leading to eruption at Santiaguito, and use our experimental results to shed light on the triggers for vulcanian explosions at Santiaguito.


40

Day 3 — Session 9 — 12.20–12.45

DECIPHERING ACOUSTIC EMISSIONS PRODUCED BY COOLING VOLCANIC ROCKS

John Browning1,2, Philip Meredith2 and Agust Gudmundsson1

1Department of Earth Sciences, Royal Holloway University of London, Egham TW20 0EX 2Department of Earth Sciences, University College London, Gower Street, London WC1E 6BT

The vast majority of published studies of thermally-induced cracking to date have focused on the generation of cracks, through detection of acoustic emissions formed during heating. This is in spite of the fact that magmatic cooling produces thermal cracks, and previous studies have indicated anecdotal evidence of increased AE during cooling of poly-crystalline rocks. As such, there is a severe need to study thermal cracking during cooling. During heating, AE is produced by cracks formed under an overall compressional regime. By contrast, cooling cracks are formed under an overall tensile regime. Therefore, both the nature and mechanism of crack formation during cooling are hypothesised to be different from those for crack formation during heating. Furthermore, it remains unclear whether cooling simply reactivates pre-existing cracks, induces the growth of new cracks, or both.

We present results from experiments based on a new method for testing ideas on cooling-induced cracking. Cored samples of volcanic rock (basaltic to dacitic in composition) were heated at varying rates to different maximum temperatures inside a tube furnace. In the highest temperature experiments samples of both rocks were raised to the liquidus temperature appropriate to their composition, forcing melt interaction and crack annealing. We present in-situ seismic velocity and acoustic emission data, which were recorded throughout each heating and cooling cycle. It is found consistently that the rate of acoustic emission is much higher during cooling than during heating. In addition, acoustic emission events produced on cooling tend to be significantly higher in energy than those produced during heating. We therefore suggest that cracks formed during cooling are significantly larger than those formed during heating. Thin-section and crack morphology analysis of our cyclically heated samples provide further evidence of contrasting fracture morphologies. These new data are important for assessing the contribution of cooling-induced damage within volcanic structures and layers such as sills and lava flows. Our observations may also help to constrain evolving ideas regarding the formation of columnar joints.


41

TALKSDAY 4


42


FAULTING, CATACLASTIC FLOW AND FRICTION OF SANDSTONES

E H Rutter1 and A Hackston1,2

1Rock Deformation Laboratory, School of Earth, Atmospheric and Environmental Sciences, University of Manchester, M13 9PL ; 2Now at Ove Arup, Lendal Arches, Tanners Moat, York YO1 6HU.

The Byerlee generalization, that to a useful approximation rock-on-rock sliding friction coefficient is independent of rock type, is widely applied in geomechanical modelling. We test this for intact and sawcut Darley Dale and Pennant sandstones (13% and 4% porosity respectively) in a standard ‘triaxial’ apparatus in both extensional and compressional loading configurations, either oven-dry or oil saturated, and over a range of pore fluid pressures. Intact samples were also faulted prior to frictional sliding. With these results it was possible to evaluate the applicability of the Mohr-Coulomb failure criterion.

For fresh faulting there was a marked difference between extensional and compressional loading, as might be expected because the intermediate principal stresses are very different. The contrast accorded with the description of failure by the modified Lade criterion. Thus a 2-dimensional Mohr-Coulomb type of description cannot capture the overall behaviour. Furthermore, the angle between maximum compressive stress and fault plane, and the apparent coefficient of friction was in all cases systematically smaller in extensional than in compressional tests. The commonly-employed assumption that fault angle can be obtained from the normal to the Mohr-Coulomb failure envelope does not hold well, neither for loading in compression nor in extension. Nevertheless, for these and other sandstones the generalization of constant friction coefficient holds quite well for a given test-type. For even more porous sandstones (e.g. Hollington sandstone, 25% porosity) the critical state description of failure must be invoked because pore collapse can dominate over microcrack-induced dilatancy at high pressures. The critical state line can be supposed to related to the friction coefficient, and as such appears, to a useful approximation, to be independent of rock type.

Frictional sliding experiments were carried out on sawcuts at angles 35, 45 and 55° to the cylindrical specimen axis. Friction coefficient for either compressional or extensional loading was not affected by the sawcut angle. The influence of pore fluid pressure on friction on 45° sawcut surfaces was investigated by bringing the sample to a confining pressure and differential stress not quite sufficient to cause sliding. Pore pressure was slowly increased until sliding started, causing some offloading, and requiring pore pressure to be increased further to maintain sliding. In this way the friction sliding line was tracked continuously all the way to low stresses. Results showed an apparent cohesive stress, attributed to failure of the pore pressure to spread over the whole slip surface (fluid fingering). Failure to develop the condition of pore pressure equals normal stress is thought to be common in nature and in the engineering of hydraulic fractures.


43

Day 4 — Session 10 — 09.50–10.15

RUPTURES PROCESSES DURING LABORATORY EARTHQUAKES

Alexandre Schubnel1, François X Passelègue1, Soumaya Latour1, Harsha S Bhat2, Stefan B Nielsen3 and Raùl Madariaga1

1Laboratopire de Géologie, ENS PARIS2Institut de Physique du Globe de Paris

3Durham University

Since the pioneering work of Brace and Byerlee (1966), the mechanism of stick-slip has been considered as a good analog to earthquake rupture propagation. Here, we report macroscopic stick-slip events in Westerly granite rock samples deformed under controlled upper crustal stress conditions in the laboratory. Experiments were conducted under triaxial loading using a high frequency acoustic monitoring array to record both particles acceleration during macroscopic stick-slip events and background microseismicity. For the first time, we also record the stress drop dynamically.

Prior to stick-slip instabilities, we observe a systematic exponential acceleration of precursory slip, but foreshocks are only observed when the normal stress becomes greater than ~50 MPa. This threshold corresponds to the normal stress above which the nucleation length becomes comparable to the size of typical fault asperities. Even in these conditions, most of precursory slip remains aseismic, but the total cumulative moment of the foreshock sequence also increases exponentially up to failure and the fault surface seem to evolve like a cascading asperity model. This exponential growth implies that the nucleation phase has a characteristic time, i.e. that both the foreshock sequence duration and precursory moment release scales with the size of the main asperity.

During the mainshock, we show that the dynamic stress drop, measured locally close to the fault plane, is almost total in the breakdown zone. The fault recovers strength within a few tens of microseconds. We demonstrate that the frictional behavior of stick-slip follows a slip-weakening law, well explained by flash heating theory, in agreement with our post-mortem microstructural analysis. Relationships between initial state of stress, rupture velocity, and stress drop demonstrate that supershear rupture is systematic at large normal stress. In these conditions, the ratio between fracture and radiated energies suggest that the rupture becomes more dispersive, meaning that the fracture energy eventually scales with both slip and stress drop. This result seems in agreement with linear elastic fracture mechanics theory and, possibly, with seismological observations.


44

Day 4 — Session 10 — 10.15–10.40

FROM LAB TO FIELD: ERUPTION FORECASTING USING VOLCANO-TECTONIC AND LOW FREQUENCY SEISMICITY

Marco Fazio1, Philip M Benson1 and Sergio Vinciguerra2,3

1Rock Mechanics Laboratory, University of Portsmouth, Portsmouth, UK.2Department of geology, university of Leicester, University Road, Leicester, LE1 7RH, United Kingdom

3British Geological Survey, Environmental Science Centre, Nicker Hill, Keyworth, Nottingham, NG12 5GG, UK

Over the last 30 years, eruption forecasting strategy based on volcanic seismicity was successfully made on several volcanoes. A quantitative method based on the fundamental law for fracture in brittle materials, corresponds in plotting the inverse rate of earthquakes versus time, revealing that the time of eruption is forecasted by extrapolating the inverse rate (y) and the time (x) in a reciprocal-rate curve plot. However, a number of known limitations exist with this method, the most serious being that it is less applicable in open-vent conditions or in volcanoes with short repose intervals (e.g. Mt Etna). The movement of volcanic fluids, such as magma, has been recognised as the source of low frequency (LF) seismicity that generally precedes an eruption, but its actual trigger and its use in forecasting methods is still under debate.

To evaluate the different kinds of volcanic earthquakes as tools for volcanic eruptions, we report new results of triaxial experiments on basalt from Mt Etna (Italy), run at different P-T-saturation conditions designed to reproduce various types of acoustic emissions (AEs), the laboratory analogue of the field scale volcanic signals. This is achieved through a servo-controlled triaxial testing machine and a state-of the-art instrumentation to record and analyse the AEs generated by fracturing in the samples and by subsequent fluid flow through the fracture network to simulate LF activity. The presence of the pore fluid has a severe influence on the characteristics of the AEs recorded during the triaxial deformation. While in dry conditions almost 10 000 events were recorded, with an exponential growth well before the sample’s failure characterized by a general decrease of the b-value, wet experiments were accompanied by about 3000 AEs, with a sudden supra-exponential growth towards the final stage and a fluctuating behaviour of the b-value. Although the failure in such conditions could be forecasted observing the decrease of the dominant frequency of the signals and the increase of the CLVD component. Forecasting the movements of fluid and its flow rate can be done considering the LF activity. Swarms of LFs only occur where the flow rate is fast enough to generate rapid pore pressure decay while the presence of very LF signals reveal the presence of a gas phase in the pore space. These results suggest that all types of events should be used to forecast volcanic eruptions, both for close- and open-vent conditions.


45

Day 4 — Session 10 — 10.40–11.05

EXTREME EVENTS DURING THE DEFORMATION OF POROUS ROCKS — FROM DISCRETE ELEMENT MODELLING TO INDUCED SEISMICITY

Ian Main1, John Greenhough1, Gergo P´al2, Ferenc Kun2 and Sabine Lennartz-Sassinek1,3

1School of Geosciences, University of Edinburgh, UK; 2Department of Theoretical Physics, University of Debrecen, Hungary

3Institute for Geophysics and Meteorology, University of Cologne, Germany

The behaviour of earthquake or acoustic emissions during loading is often studied by parameters representing the whole population of events — such as the temporal evolution of event rate, or the scaling of frequency with inter-event time or event magnitude. However, in many applications the important phenomena are the extreme or record-breaking events — for example the magnitude of induced or triggered earthquakes, especially if they are large enough to be felt. A key question is the effect of the partitioning between seismic and aseismic strain on the seismic potential of extreme events. First we show from published data that the strain partition coefficient increases only slowly with respect to injected volume, at a decelerating rate. We then investigate the statistics of record-breaking rupture events generated by computer simulations of the uniaxial compression of cylindrical samples in a discrete element model. The number of records grows initially as a decelerating power law of the number of events, followed by a bifurcation to an accelerating pattern immediately prior to failure. The lifetime of records is power-law distributed with a relatively low exponent. The results show two clearly separate regimes in the time evolution: initially the statistics are consistent with sampling of a random process, most likely due to the dominance of structural disorder in the sub-critical stage of loading. Record breaking then accelerates as macroscopic failure is approached, when near-critical physical interactions and spatial and temporal correlations dominate. In comparing the simulations with induced seismicity data, the scaling of extreme events in induced seismicity is inconsistent with sub-critical processes. The results of the simulations can be used in principle as a benchmark for interpreting the temporal evolution of such extreme events for a variety of applications, from non-destructive testing of geological materials to risk management for induced seismicity.


46

Day 4 — Session 11 — 11.30–11.55

THE ROCK PHYSICS OF FIBER-REINFORCED ROCKS HELPS EXPLAIN UPLIFTS AT CAMPI FLEGREI CALDERA

T Vanorio1 and Waruntorn Kanitpanyacharoen1,2

1Stanford Rock Physics Laboratory (SRPL), Department of Geophysics, Stanford University, USA2Currently at the Department of Geology, Faculty of Science, Chulalongkorn University, Thailand

The caldera of Campi Flegrei is one of the active hydrothermal systems of the Mediterranean region experiencing notable unrest episodes in a densely populated area. During the last crisis of 1982–1984, nearly 40 000 people were evacuated for almost two years from the main town of Pozzuoli, the Roman Puteoli, due to the large uplifts (~2 m over two years) and the persistent seismic activity. The evacuation severely hampered the economy and the social make-up of the community, which included the relocation of schools and commercial shops as well as the harbour being rendered useless for docking. Despite the large uplifts, the release of strain appears delayed. Seismicity reaches a magnitude of 4.0 only upon relatively large uplifts (~70–80 cm) contrary to what is generally observed for calderas exhibiting much lower deformation levels. Over and above the specific mechanism causing the unrest and the lack of identification of a shallow magmatic reservoir (<4 km) by seismic data, there is a core question of how the subsurface rocks of Campi Flegrei withstand a large strain and have high strength. We combine high-resolution microstructural and mineralogical analyses with the elastic and mechanical properties of well cores from the deep wells of San Vito (SV1 and SV2) and Mofete (MF1, MF2, MF5) that were drilled in the area right before the unrest of 1982–1984. The rock physics analysis of the well cores provides evidence for the existence of two horizons, above and below the seismogenic area, underlying a natural, coupled process. The basement is a calc-silicate rock housing hydrothermal decarbonation reactions, which provide lime-rich fluids. The caprock above the seismogenic area has a pozzolanic composition and a fibril-rich matrix made of intertwining filaments of ettringite and tobemorite, resulting from lime-pozzolanic reactions.

These findings provide evidence for a natural process reflecting that of the engineering of the mortar of the Roman concrete. The formation of fibrous minerals by intertwining filaments confers shear and tensile strength to the caprock, contributing to its ductility and increased resistance to fracture. The importance of the findings reported in this study lies not only on the fibrous and compositionally nature of the caprock but also on its possible physicochemical deterioration. Given the P-T-XCO2 conditions regulating the decarbonation reactions, the influx of new brine into the Campi Flegrei system lowers the temperature of the decarbonation reaction and dilutes the existing CO2, thus triggering additional CO2, methane, and steam to form. As these gases rise toward the surface, they are halted by the natural cement layer, leading to pore pressure increase and subsequent ground deformations.


47

Day 4 — Session 11 — 11.55–12.20

X-RAY COMPUTED TOMOGRAPHY INVESTIGATION OF STRUCTURES IN CLAYSTONES FROM LARGE SCALE TO SMALL SCALE

A Kaufhold1, W Gräsle1, H Halisch2 and G Zacher3

1Federal Institute for Geosciences and Natural Resources, Stilleweg 2, 30655 Hannover, Germany2Leibniz Institute for Applied Geophysics, Stilleweg 2, 30655 Hannover, Germany

3GE Sensing & Inspection Technologies GmbH, phoenix|x-ray, Niels-Bohr-Str. 7, 31515 Wunstorf, Germany

In the past years, X-ray Computed Tomography (CT) has become more and more widely used in geoscience for investigations from the micro scale (e.g.microfossils) up to the decimetre scale (e.g. drill cores or soil columns). The method is (amongst others) applied for structural investigations, in situ investigations of processes during tests, or for the study of porosity and permeability analysis for sediments and rocks. Hence a variety of systems were adapted to these applications and the investigated specimen size.

In the present study, we present CT results from an Opalinus Clay core (diameter ~100 mm), originating from Underground Rock Laboratory (URL), Mont Terri. Detailed information before and after the mechanical testing are required for the understanding of deformation processes during mechanical testing. Thus, it is necessary to obtain micro fabric information of the undisturbed specimen and an overview of the deformed specimen after the mechanical test (Figure 1a). Using the overview scans it is possible to select specific regions of interest (ROIs) which can be analysed more in detail by using high resolution CT devices. The suitable position for further micro plugging can be identified from fast 3D scans (Figure 1b).

Selected areas are analysed by using high-resolution CT techniques as well as mineralogical and geochemical methods. The overall aim of the investigation of the Opalinus Clay (LT-A Project, Mont Terri) is to understand the rock deformation processes upon mechanical stressing. This behaviour is largely governed by the microstructure. CT investigations, therefore, are key methods for visualizing and understanding the processes during the mechanical tests.

Additionally, chemical and mineralogical methods are used not only to characterize the material but also to identify homogeneous areas which can be considered to be representative for the entire rock. Hence, the CT information gathered from a small volume can be used to understand the mechanical processes on a large scale.


48


49

POSTERS SESSION 1


50

Poster session 1 — Board 01

LABORATORY PERMEABILITY AND SEISMIC VELOCITY ANISOTROPY MEASUREMENTS ACROSS THE ALPINE FAULT, NEW ZEALAND

M Allen1, D Tatham1, D Faulkner1 and E Mariani1

1School of Environmental Sciences, University of Liverpool, United Kingdom

The Alpine Fault zone is transpressional plate boundary running along the western edge of the New Zealand’s South Island. The Deep Fault Drilling Project conducted two scientific bore holes upon the Alpine Fault zone, near Whataroa, South Island. Each borehole reached respective depths of ~100 m and ~151 m allowing the collection of a suite of fault rock lithologies, ranging from ultramylonites, ultracomminuted gouges and variably foliated and unfoliated cataclasites from both the hanging wall and footwall of the Alpine Fault. Drilling revealed a typical shallow fault zone structure, localised principal slip zones of gouge nested within a damage zone comprised of cataclasites and ultramylonites.

A suite of laboratory experiments has been undertaken in order to determine the permeability and seismic wave velocities at incremental depths across the Alpine Fault at shallow depths. Tested specimens were taken incrementally to maximum effective pressures of 105 MPa with Pore fluid pressures maintained at ~5 MPa.

Using the Pulse Transient technique permeability measurements at effective pressures of 5 MPa range from 10–17 to 10–20 m2 decreasing to 10-18 to 10-21 m2 at effective pressures of 105 MPa.

Seismic velocity measurements were performed up to 105 MPa effective pressures with P wave velocities ranging from ~4500 to 5900 m/s and S wave velocities ranging from 2300 to 3600 m/s.

Each specimen was tested in three orthogonal directions, orientated according to original core orientation markers or when present to foliation. Testing in multiple orientations has revealed that some permeability and seismic velocity anisotropy exists, likely caused through the variation of microstructures with depth. These microstructural features include foliation, clay and carbonate filled fractures and according to density and orientation relative to the tested directions.


51


BULK MODULUS DISPERSION AND ATTENUATION OF A FLUID-SATURATED LAVOUX LIMESTONE

J Borgomano1, L Pimienta1, J Fortin1 and Y Guéguen1

1Laboratoire de Géologie, Ecole Normale Supérieure, Paris, FR.

Frequency effects are known to affect the elastic properties of porous materials when saturated with fluids. Therefore, fluid-saturated rocks are dispersive and interpreting seismic field data from laboratory high-frequency measurements (~1 MHz) cannot be straightforward. A new experimental method has been developed at the Ecole Normale Supérieure of Paris to measure the bulk modulus of a rock’s sample at both low (f ~10-3–100 Hz) and high (f ~1 MHz) frequencies, to investigate these frequency effects. This apparatus has been calibrated on standard materials (Glass, Plexiglas and Gypsum) and several measurements on well known Fontainebleau sandstones have proven its reliability.

Dispersion and attenuation effects in porous rocks can be discussed in terms of three fluid-flow regimes: drained, undrained and unrelaxed. The transitions between these regimes are characterized by cut-off frequencies around which the porous medium is highly affected by dispersion and attenuation. These cut-off frequencies depend on different parameters such as the fluid’s viscosity , the intrinsic permeability k, the microstructure’s average crack aspect ratio , the drained bulk modulus Kd or the length of the sample L.

Carbonate rocks are characterized by complex microstructures and heterogeneous pore types. Analyzing the frequency effects in terms of microstructure/fluid systems is therefore more challenging than in a homogeneous rock type. In this paper we present our first bulk modulus and attenuation measurements, using oscillation tests and PZT sensors, on a hydrostatically stressed pure calcite oolitic limestone from Lavoux, at different frequencies f (low : [0.004–0.4] Hz; ultrasonic: 1 MHz) and effective pressures (2.5–30 MPa), under dry, water-saturated and glycerin-saturated conditions. The Lavoux limestone presents a dual porosity distribution (macro- and microporosity). The sample’s total porosity is around 24% and its permeability is about 10 mD.

Results show that dispersion is only observed under glycerin-saturated conditions, correlated with a peak in attenuation. According to the cut-off frequency, this corresponds to the drained/undrained transition.

Figure: Low frequency measurements of the Lavoux' bulk modules, under different pressure and saturating conditions.


52


HIGH-VELOCITY FRICTIONAL PROPERTIES OF FAULT CORE GOUGES AND CATACLASITES, ALPINE FAULT, NEW ZEALAND

Carolyn Boulton1,2*, Lu Yao3, Daniel R Faulkner2, Shengli Ma3, Toshihiko Shimamoto3, John Townend4, Virginia G Toy5 and Rupert Sutherland6

1Department of Geological Sciences, University of Canterbury, Christchurch, New Zealand2School of Environmental Sciences, University of Liverpool, Liverpool, United Kingdom

3State Key Laboratory of Earthquake Dynamics, Institute of Geology, China Earthquake Administration, Beijing, China4School of Geography, Environmental, and Earth Sciences, Victoria University of Wellington, Wellington, New Zealand

5Department of Geology, University of Otago, Dunedin, New Zealand6GNS Science, Lower Hutt, New Zealand

The Alpine Fault in New Zealand is a major plate-bounding structure that typically slips in ~M8 earthquakes every c.330 years. To investigate the near-surface, high-velocity frictional behavior of surface- and borehole-derived Alpine Fault gouges and cataclasites, 19 rotary shear experiments were conducted at 1 MPa normal stress and 1 m/s equivalent velocity. Experiments were first conducted on room-dry fault rocks, and then repeated with the addition of 25 wt.% deionized water to simulate fluid-saturated conditions. In the room-dry experiments, the peak coefficient of friction (µp=p/n) of Alpine Fault cataclasites and fault gouges was consistently high (mean µp=0.65±0.10). Variations in mineralogy and permeability were more apparent during the wet experiments, wherein the peak coefficient of friction of the phyllosilicate-poor cataclasites (mean µp=0.63±0.06) was higher than that of the fault gouges (mean µp=0.27±0.21). All fault rocks exhibited very low steady state coefficients of friction (µss) (room-dry experiments mean µss=0.18±0.04; wet experiments mean µss=0.09±0.04). The five experiments conducted on wet smectite-bearing principal slip zone fault gouges yielded the lowest peak friction coefficients (µp=0.13–0.21), the lowest steady state friction coefficients (µss=0.02–0.10), and, commonly, the lowest breakdown work values (WB=0.07–0.57 MJ/m2) of all the experiments performed. Microstructural interpretations, combined with axial displacement data, indicate that thermal pressurization of ambient pore fluid, dehydrated adsorbed water and/or dehydrated smectite interlayer water was the primary dynamic-weakening mechanism responsible for low friction in the high velocity experiments. Given sufficient acceleration, earthquake rupture propagation through the smectite-bearing PSZ fault gouges is energetically favourable over the cataclasites.


53


SUBSIDENCE ABOVE A GAS RESERVOIR AS AFFECTED BY FLOW OF ROCKSALT (THE CAP ROCK)

G Marketos1, R Govers1 and C J Spiers1

1Department of Geoscience, Utrecht University, The Netherlands

Production from gas reservoirs leads to compaction at the reservoir level and ground displacements that induce surface subsidence. Understanding the patterns of this subsidence can be very important, especially for low-lying coastal regions. Here we focus on gas reservoirs that are capped by rocksalt, a material that behaves elastically over short timescales but then flows over larger timescales so as to relax the shear stresses induced in it. We create simplified representations of the salt-capped reservoir in a Finite Element model and observe that rocksalt flow can induce time-varying subsidence. The mechanisms behind this time-varying subsidence will be presented. It will be shown that rocksalt flow can induce significant subsidence above a gas reservoir (see e.g. [1]), and so should be modelled as accurately as possible. This contribution will also briefly discuss the various constitutive models available for rocksalt flow and the uncertainties when choosing one for a specific application.

References:G Marketos, R Govers, C J Spiers (2015): Ground motions induced by a producing hydrocarbon reservoir that is overlain by a viscoelastic rocksalt layer: a numerical model. Geophysical Journal International doi: 10.1093/gji/ggv294


54


DETECTION OF MOVING CAPILLARY FRONT IN POROUS ROCKS USING X-RAY AND ULTRASONIC METHODS

C David1, D Bertauld1,2, J Dautriat2, J Sarout2, B Menendez1 and B Nabawy3

1Department of Geosciences and Environment, University Cergy-Pontoise, France2CSIRO Energy, Perth, Western Australia

3National Research Center, Department of Geophysical Sciences, Cairo, Egypt

Several methods are compared for the detection of moving capillary fronts in spontaneous imbibition experiments where water invades dry porous rocks. These methods are:

(i) the continuous monitoring of the mass increase during imbibition, (ii) the imaging of the water front motion using X-ray CT scanning, (iii) the use of ultrasonic measurements allowing the detection of velocity, amplitude and spectral content of the

propagating elastic waves,(iv) the combined use of X-ray CT scanning and ultrasonic monitoring.

It is shown that the properties of capillary fronts depend on the heterogeneity of the rocks, and that the information derived from each method on the dynamics of capillary motion can be significantly different. One important result from the direct comparison of the moving capillary front position and the P wave attributes is that the wave amplitude is strongly impacted before the capillary front reaches the sensors, in contrast with the velocity change which is concomitant with the fluid front arrival in the sensors plane. Using a leading edge CT scanner allowed us to get 3D images of the volume invaded by the water during imbibition and to image the geometry of the moving capillary front at different stages.


55


THE KG2B PROJECT: A WORLD-WIDE BENCHMARK OF LOW PERMEABILITY MEASUREMENT

C David1, J Wassermann1, F Amann2 and the KG2B Team3

1University Cergy-Pontoise, France2ETH Zürich, Switzerland

3many people from 24 other labs around the world

Following a workshop on « The challenge of studying low permeability materials » held in Cergy-Pontoise University in December 2014, a benchmark of low permeability measurements has been proposed to the attendees. A total of 26 laboratories from 8 different countries throughout the world volunteered and are participating to the benchmark. The contribution of each team will be to measure the permeability of a common low permeability material using their preferred technique. A wide range of different methods were proposed (i) direct measurements using steady-state, transient or oscillatory flow techniques, (ii) indirect permeability estimation using proxys, and (iii) numerical simulations using information on microstructural properties.

Several options have been explored for the material selection, and finally we decided to choose a hard rock, the Grimsel granodiorite (Switzerland), within the framework of an exciting scientific project of our Swiss colleagues. The benchmark was therefore called the 'KG2B' project, which means 'K for Grimsel Granodiorite Benchmark'. Fresh cores from the Swiss Grimsel test site, an underground research laboratory in hard rock will be drilled mid-August 2015. The drilling of these cores is part of a scientific project funded through the Swiss Competence Center of Energy Research — Supply of Electricity (SCCER-SoE). The aim of the SCCER-SoE is to perform a series of demonstration experiments on various scales (up to 1 km) that allow implementing deep geothermal energy in Switzerland. Beside other experiment the SCCER-SoE is currently executing a large-scale stimulation and circulation in situ experiment at the Grimsel Test Site. The experiment requires the drilling of approximately 600 m of boreholes which will be used as 1) injection boreholes and 2) monitoring boreholes. All boreholes will be cored using single- or double tube core barrels in the Grimsel Granodiorite. The cores retrieved from the boreholes will be cut at a length of about 100 mm. The porosity of the Grimsel Granodiorite is about 0.7%, and the permeability should range between 0.1 and 1 µD. The benchmark study is of great value for both the stimulation and circulation experiment which involves a set of fault zones embedded in a low permeability rock matrix.

Before sending the blocks to the participants, a quality checking will be done by estimating the reproducibility and anisotropy of the material through measurements of P wave velocity. A 'check-list' with basic instructions to be followed and a document in which each participant will provide detailed information on his measurement protocol will be provided. The results of the benchmark will be compiled, analyzed and hopefully published before the end of this year. In the future we plan to organize a second round of the benchmark focusing this time on synthetic materials with controlled properties and microstructure.


56


TOWARDS NEUTRON IMAGING OF COUPLED DEFORMATION AND FLUID FLOW IN POROUS, GRANULAR ROCKS

E Tudisco1, S A Hall1,2, S D Athanasopoulos1 and J Hovind3

1Division of Solid Mechanics, Lund University, Lund Sweden2European Spallation Source AB, Lund, Sweden

3Paul Scherrer Institute, Villigen, Switzerland

In this work, 'in-situ' triaxial testing with neutron imaging is used to quantify the evolution of deformation in an artificial porous, granular rock coupled to the evolution of fluid flow through the sample (in-situ meaning that tests are carried out within the neutron imaging station). Previously we have demonstrated the potential of neutron imaging to study rock deformation and fluid flow through imaging of deformed rock samples (Hall et al., 2013; Tudisco et al, 2015a;b) with pre-/post-mortem tomography, Digital Volume Correlation (DVC) and quantified radiography of fluid flow. These results will be reviewed briefly before presenting new work on in-situ triaxial testing with time-lapse neutron tomography in combination with DVC-based 3D strain field mapping to follow the 4D evolution of the deformation. Furthermore, analysis of fluid flow evolution through the same specimens at different stages of deformation was possible using neutron radiography of water flooding into the heavy-water saturated specimen. This coupled full-field deformation and flow mapping approach can lead to new insights into how fluid-flow properties are changed in the regions where the most significant deformation occurs, i.e., in localised deformation zones, as opposed to measuring just bulk changes. Furthermore, in-situ experiments enable the mechanical and fluid-flow evolution, plus their coupling, to be monitored (see figure), under realistic conditions and without removing the sample from the cell.

1. Hall, S A. 2013, Geophysical Research Letters, 40, 2613–2618.

2. Tudisco, E, Hall, S A, Hovind, J, Khardjilov, N, Charalampidou, E-M and Sone, H. 2015, , Physics Procedia, in press.

3. Tudisco, E, Hall, S A, Hovind, J, Khardjilov, N, Charalampidou, E-M and Sone, H. 2015, in Proceedings of the South American Congress On Rocks Mechanics, in press.


57


THE EFFECT OF LONG-TERM FLUID-ROCK INTERACTIONS ON THE MECHANICAL PROPERTIES OF RESERVOIR ROCK — A CASE STUDY OF THE

WERKENDAM NATURAL CO2 ANALOGUE FIELD

Suzanne Hangx1,2, Pieter Bertier3, Elisenda Bakker2, Georg Nover4 and Andreas Busch1

1Shell Global Solutions, Kesslerpark 1, 2288 GS Rijswijk, the Netherlands2Department of Earth Sciences, Utrecht University, Budapestlaan 4, 3584 CD Utrecht, the Netherlands

3Clay and Interface Mineralogy, RWTH Aachen University, Intzestrasse 1, 52056 Aachen Germany4Steinmann Institute, Bonn University, Poppelsdorfer Schloss, 53115 Bonn, Germany

Geological storage of CO2 is one of the most promising technologies to rapidly reduce anthropogenic emissions of carbon dioxide. During long-term geological storage of CO2, fluid-rock interactions, induced by the formation of carbonic acid, may affect the mineralogical composition of the reservoir rock. Commonly expected reactions include the dissolution of carbonate and/or sulphate cements, as well as the reaction of primary minerals (feldspars, clays, micas) to form new, secondary phases. In order to ensure storage integrity, it is important to understand the effect of such fluid-rock interactions on the mechanical behaviour of a CO2 storage complex. However, most of these reactions are very slow, which limits the ability to study coupled chemical-mechanical processes in the lab. A possible way to circumvent long reaction times is to investigate natural CO2 analogue fields, which experienced CO2-exposure for thousands of years.

In this study, we looked at the Dutch Werkendam natural CO2 field and its unreacted counterpart (Röt Fringe Sandstone, Werkendam, the Netherlands). We focussed on CO2-induced mineralogical and porosity-permeability changes, and their effect on mechanical behaviour of intact rock. Overall, CO2-exposure did not lead to drastic mineralogical changes, though markedly different porosity-permeability relationships were found for the unreacted and exposed material. The limited extent of reaction was in part the result of bitumen coatings protecting specific mineral phases from reaction, as well as from low water-to-rock ratios. In local, mm-sized zones, significant anhydrite dissolution was observed, most likely due to higher pre-exposure porosity of these zones, resulting in higher water-to-rock ratios. For most of the reservoir the long-term mechanical behaviour after CO2-exposure could be described by the behaviour of the unreacted sandstone, while these more ‘porous’ zones were significantly weaker. The lower strength of these zones was in part the result of the initial (higher) porosity, but also in part caused by the dissolution of the grain-cementing anhydrite. However, to distinguish between which of these two factors played a more significant role on rock strength is difficult.


58


MECHANICAL PROPERTIES OF PROTOLITH AND GOUGE FROM THE CARBONERAS FAULT, SE SPAIN: HOW MECHANICS INFLUENCE

FAULT ZONE STRUCTURE

A Kätker1, D Faulkner2, H Leclère2, C Boulton2, J Renner1

1Ruhr University Bochum, Institute of Geology, Mineralogy and Geophysics, Bochum, Germany2Rock Deformation Laboratory, Department of Earth and Ocean Sciences, University of Liverpool, Liverpool, UK

The Carboneras fault in southeastern Spain is a 1 km-wide fault zone with ~40 km of strike-slip offset. The fault zone displays multiple strands of fault gouge and fractured lenses of protolithic mica schist. To understand why the fault zone developed multiple slip surfaces, we conducted friction tests on fault gouge and triaxial compression tests on cores of the protolith using a triaxial apparatus at room temperature. For the gouge, we performed normal-stress stepping experiments (15-30-60-90-100 MPa), as well as velocity stepping experiments (0.3-3-0.3-3-0.3-3 µm/s) at 50 MPa effective normal stress (n’). To realize that, the wet experiments were conducted at confining pressures of 25, 40, 70, 100 and 110 MPa and a constant pore pressure of 10 MPa, which was also used for the velocity stepping experiments. Samples slid up to 4 mm in the direct shear configuration. Results show friction coefficients (µ=/n’) between 0.45–0.66 (dry) and 0.31–0.37 (wet) and clear velocity strengthening behavior. Dry cores of protolith were triaxially compressed at confining pressures of 20, 50, 80 and 110 MPa. We loaded the samples to failure and then recorded their residual strength, sliding them for 4 mm as well. The differences between the peak strength and frictional strength are between 5 and 25%, with little clear systematics to confining pressure. The cores typically developed conjugate shear fractures with angles of approximately 45° to the sample axes. The frictional strength of these conjugate shear fractures corresponds to friction coefficients between 0.48–0.62 (dry), values quite similar to that of the gouge. We prepared thin sections of all samples for microstructural analyses. The similarity in strength of dry protolith and gouge provides an explanation for the wide fault zone with multiple strands. Our mechanical results and structural observations on sample scale indicate that it was almost as easy to develop new fault planes, as to continue sliding on pre-existing slip surfaces.


59


SIMULATION OF HYDROCARBON PRIMARY MIGRATION: MULTI-EXPERIMENTAL APPROACH

M Kobchenko1, A Pluymakers1, D Dysthe1, F Renard1,2 and A-M Sørenssen1

1Department of Physics, University of Oslo, Norway2ISterre, Univ. Grenoble Alpes, France

For a consistent study of the prospects for the immature organic-rich shales as a source of hydrocarbons, the natural phenomenon of primary migration needs to be understood. The oil and gas generated during maturation in the low permeability organic-rich shale migrate via two simultaneous processes: diffusion through the rock matrix and pore-pressure induced fracturing. Depending on the relation between rates of diffusion and hydrocarbon production, networks of different densities can be produced. Interpreting the results of an analogue experiment with gelatine used as the shale analogue and yeast producing CO2 — as the hydrocarbon source, we put forward a hypothesis suggesting an intermittent character of oil and gas escape process through the generated fracture networks. We investigate its implications for primary migration using time-resolved 3D imaging of shales during combustion and simulation of reservoir conditions in high-pressure/high-temperature autoclaves.


60


WHAT DO FRACTURES IN SHALE LOOK LIKE AT DEPTH? A STUDY ON THE ROUGHNESS OF THE VEIN-ROCK INTERFACE VS. OPEN FRACTURES

A Pluymakers1, M Kobchenko1, D Dysthe1 and F Renard1,2

1Department of Physics, University of Oslo, Norway2ISterre, Univ. Grenoble Alpes, France

The possibility of enhanced gas recovery from methane-bearing shale rocks receives currently a lot of attention. How to determine the permeability of these rocks, and how it is controlled, i.e. by pre-existing or induced fractures, represents a key challenge. One of the principal questions here is how to determine which fractures are present in-situ, and which fractures originate from sample recovery. One of the ways to characterize fractures is through their roughness. In this study, we determine fracture roughness and geometry both of veins and decompression fractures of the Pomeronian shale from Poland, recovered from ~4 km depth. Since vein formation occurs at depth, the topography of the vein-rock interface is preserved even when samples are extracted from the subsurface. Here, we have imaged a series of shale sample in 3D using X-ray microtomography performed on a laboratory tomograph and on the beamline ID19 at the European Synchrotron Radiation Facility. Several voxel spatial sizes were used, in the range 0.2 to 20 micrometers, allowing a multi-scale analysis of these rocks. We also compare fracture and veins roughness values from tomography scans to results obtained using white light interferometry, where these different measurements provides us with information on how scale, resolution and method affect roughness values. Preliminary results indicate low resolution CT scans show a difference between the roughness of the vein-rock interface and decompression fractures. The results from this study provide input for (reactive) flow models, to study how efficient gas recovery can be.


61


3D PRINTING CARBONATE MICROSTRUCTURES: PRELIMINARY POROSITY-PERMEABILITY TRENDS WITH APPLICATIONS TO THE

DECARBONATION REACTION

D Head1 and T Vanorio1

1Department of Geophysics, Stanford University, USA

There has long been a dichotomy between two of the main approaches used in rock physics analysis — digital and experimental. In the laboratory we can explore still unknown frontiers of rock behaviour by controlling pressure, temperature, fluid properties, and time. We can carefully select natural samples or manually create synthetic samples to control initial mineralogy, porosity, permeability, sorting, and average grain size. However, once the natural or synthetic sample exists, we cannot manually change or exactly reproduce it. On the other hand, in digital rock physics each sample and experiment is fully reproducible. Minute, pore scale details can be observed and manipulated. However, the utility of the solution relies heavily on both input resolution and a priori knowledge of the system used to inform the boundary conditions.

Bridging this gap has become more important than ever as we add layers of complexity to both models and experiments in an attempt understand the coupled thermo-chemo-mechanical changes controlling transport and elastic properties of carbonate diagenesis. The advent of modern 3D printing has provided an unprecedented opportunity to combine the strengths of both experimental and digital rock physics. Through experimental diagenesis, we can study the evolution of properties resulting from the dynamic changes in rock systems exposed to reactive conditions. These results can be used to inform the boundary conditions of digital rock models. These models can be iteratively changed, printed, and measured to describe the effect of diagenetically induced changes in the rock microstructure on the bulk acoustic and transport properties.

In this study, we take a two-pronged approach. First, we test the feasibility of measuring 3D prints of natural carbonate pore geometries in the laboratory. We investigate the effect of changing the size of a specific natural carbonate pore geometry on the frame independent properties porosity and permeability and compare the laboratory measurements to the results of numerical simulations. These preliminary tests show that it is possible to use an iterative, grain-scale geometry modification and measurement workflow that utilizes 3D printing. Second, we induce the decarbonation reaction in a carbonate deposit injected with hot silicate-bearing fluids, in a temperature-pressure space not previously explored. These results show that we can quantify changes to the acoustic and transport properties of the sample when exposed to those diagenetic conditions.

Ultimately we will use a workflow designed to iteratively combine baseline CT-scanned rock volumes, experimentally-driven modification of the digital rock volumes, and measurements on printed rock models to test grain-scale change hypotheses on bulk sample properties.


62


THE ARCHITECTURE AND FRICTIONAL PROPERTIES OF FAULTS IN SHALE

Nicola De Paola1, Rosanne Murray1, Mark Stillings1, Jonathan Imber1 and Robert Holdsworth1

1University of Durham, Rock Mechanics Laboratory, Earth Sciences, Durham, United Kingdom

The geometry of brittle fault zones and associated fracture patterns in shale rocks, as well as their frictional properties at reservoir conditions, are still poorly understood. Nevertheless, these factors may control the very low recovery factors (25% for gas and 5% for oil) obtained during fracking operations.

Extensional brittle fault zones (maximum displacement <3 m) cut exhumed oil mature black shales in the Cleveland Basin (UK). Fault cores up to 50 cm wide accommodated most of the displacement, and are defined by a stair-step geometry, controlled by the reactivation of en-echelon, pre-existing joints in the protolith. Cores typically show a poorly developed damage zone, up to 25 cm wide, and sharp contact with the protolith rocks. Their internal architecture is characterised by four distinct fault rock domains: foliated gouges; breccias; hydraulic breccias; and a slip zone up to 20 mm thick, composed of a fine-grained black gouge. Hydraulic breccias are located within dilational jogs with aperture of up to 20 cm, composed of angular clasts of reworked fault and protolith rock, dispersed within a sparry calcite cement.

Velocity-step and slide-hold-slide experiments at sub-seismic slip rates (microns/s) were performed in a rotary shear apparatus under dry, water and brine-saturated conditions, for displacements of up to 46 cm. Both the protolith shale and the slip zone black gouge display shear localization, velocity strengthening behaviour and negative healing rates. Experiments at seismic slip rates (1.3 m/s), performed on the same materials under dry conditions, show that after initial friction values of 0.5–0.55, friction decreases to steady-state values of 0.1–0.15 within the first 10 mm of slip. Contrastingly, water/brine saturated gouge mixtures, exhibit almost instantaneous attainment of very low steady-state sliding friction (0.1).

Our field observations show that brittle fracturing and cataclastic flow are the dominant deformation mechanisms in the fault core of shale faults, where slip localization may lead to the development of a thin slip zone made of very fine-grained gouges. The velocity-strengthening behaviour and negative healing rates observed during our laboratory experiments, suggest that slow, stable sliding faulting should take place within the protolith rocks and slip zone gouges. This behaviour will cause slow fault/fracture propagation, affecting the rate at which new fracture areas are created and, hence, limiting oil and gas production during reservoir stimulation. During slipping events, fluid circulation may be very effective along the fault zone at dilational jogs — where oil and gas production should be facilitated by the creation of large fracture areas — and rather restricted in the adjacent areas of the protolith, due to the lack of a well-developed damage zone and the low permeability of the matrix and slip zone gouge. Finally, our experiments performed at seismic slip rates show that seismic ruptures may still be able to propagate in a very efficient way within the slip zone of fluid-saturated shale faults, due to the attainment of instantaneous weakening.


63


PRESSURE-SENSITIVITY OF PERMEABILITY OF SHALES, AND IMPLICATIONS FOR SHALE GAS RESERVOIR EVALUATION AND PRODUCTION

R L Taylor1, J. Mecklenburgh1, R McKernan1 and E H Rutter1

1School of Earth, Atmospheric and Environmental Sciences, University of Manchester, UK

Exploitation of shale gas using hydraulic fracture, or the performance of a shale as a seal above a conventional gas reservoir depends on the rate at which gas can flow through the pores of the rock matrix. Permeability of shale to gas depends on the difference between gas pressure in the intergranular pore spaces and the externally applied overburden pressure, but modelling of reservoir/seal behaviour depends upon knowing the form of the permeability/pressure relationship and the values of the parameters. These can only be determined through laboratory measurements.

Shale permeability is not constant, but varies significantly with confining and pore pressures and those tested exhibit common generic patterns. We seek to establish whether these patterns of permeability behaviour exist for different shales by determining their parameters experimentally. Pressure sensitivities of the gas permeability of two mineralogically comparable Jurassic shales were compared over the whole range of reservoir pressure conditions.

(1) Whitby shale — collected from the low-tide level at Runswick bay, North Yorkshire, England, so that the rock was always submersed in sea water. Clay-bearing, well-foliated silt-rich mudstone, porosity 8%, total organic carbon (TOC) 1.5%. Samples were measured dried to constant weight at 60°C.

(2) ‘Texas’ shale — core samples provided by BG International, depth and precise location unspecified. The section of core used was chosen for its homogeneity but still exhibits some fractures. Silty-argillaceous to silty-calcareous, clay-bearing mudstone, well-foliated, porosity 9%, TOC 4%. Samples were measured dried to constant weight at 60°C and as-supplied. As-supplied water saturation is 40%+10%.

The relationship between permeability, total hydrostatic pressure, and pore fluid pressure follows the general law

k = A exp(-(Pc–Pp)) (1 + D / Pp)

in which A, , and D are empirical parameters. and describe the sensitivity to confining and pore pressures. To determine parameters and it is necessary to carry out experiments varying confining pressure at constant Pp and varying pore pressure at constant Pc.

Permeability can vary by more than 3 orders of magnitude over the whole reservoir pressure range. Slip (Klinkenberg) flow is only significant at gas Pp < 5 MPa. Partial saturation of pore space with water or other fluids leads to a reduction in k, and in all cases flow is highly anisotropic. It is important to confirm that the rock responds elastically to changes in load when making permeability measurements. Over the range of reservoir pressure conditions pore volume changes with pressure are elastic and recoverable, except during the initial application of pressure.

If pressure sensitivity of permeability is not taken into account, reservoir evaluations from well tests will lead to substantial overestimation of original gas in place and of likely yield with time.


64


EFFECT OF CLAY MINERALS ON THE EFFECTIVE PRESSURE LAW IN CLAY-RICH SANDSTONES

W L Xiao, L Jiang, M Li, J Z Zhao and L L Zheng

1State Key Laboratory of Oil and Gas Reservoir Geology and Exploitation, Southwest Petroleum University, China * W. L. Xiao. [[email protected]]

The relative sensitivity of permeability to pore fluid pressure Pp and confining pressure Pc for rocks can be expressed as the effective pressure coefficient nk. To further study the effect of clay minerals on nk, permeability data were measured under conditions of lowering Pp in different constant-confining pressures for clay-rich sandstones, and then the effective pressure coefficients were calculated using the modified slide method. We found that the values of nk were constant, which was similar to previous findings. Moreover, the values of nk in I/S-sandstones and chlorite-sandstone agreed with the clay shell model, and the values of nk in previous kaolinite-sandstones were consistent with the clay particle model. However, the nk values of our kaolinite-sandstones are less than 1.0, which is different from previous observations (nk >1.0, even more than 1.0). Finally, it was found that the coefficient nk is affected by the type of clay mineral, the preparation of the samples, and the type of the pore fluid.


65


VOLCANIC PERMEABILITY: A HIGH TEMPERATURE EXPERIMENTAL INSIGHT INTO SILICIC MAGMA DEGASSING

A Chadderton1, P Sammonds1, P Meredith2, R Smith1 and H Tuffen3

1Institute for Risk and Disaster Reduction, University College London, UK2Department of Earth Sciences, University College London, UK

3Lancaster Environment Centre, Lancaster University, UK

Experimentally determined permeability observations have provided the basis for numerous theories of magmatic degassing. Recent enhancements to the High Temperature Triaxial Deformation Rocchi Cell at UCL (HTTDC) (Rocchi et al., 2004) have enabled us to make permeability measurements on 25mm x 75mm core samples at both elevated temperature and hydrostatic pressure (Gaunt et al, 2013). Specifically, we present the results of several suites of permeability data on samples of dome dacite from the 2004–08 eruption of Mount St Helens, and rhyolite collected from the lava dome formed during the 2008–2010 eruption of Chaitén, Chile. All permeability measurements were conducted under an effective pressure of 5 MPa (confining pressure of 10 MPa and pore fluid pressure of 5 MPa) and temperatures up to 900°C, using the steady-state flow technique. Samples were cooled to room temperature between each high temperature test, and the permeability of each sample was re-measured before heating to the next temperature increment in the series. The results show that the permeability of each sample decreased by approximately 3 orders of magnitude up to 600°C. The room temperature permeability after each heating cycle also decreased initially, but after heating the sample to 300°C the room temperature permeability began to recover, though not fully to its original state. Most recently, the capabilities of the HTTDC apparatus have been further extended to enable permeability measurements to be made during triaxial deformation of test samples under similar temperature and pressure conditions. Initial results from this entirely new methodology will also be presented.

These new experimental results are being applied to enhance our understanding of the complex issue of silicic magma degassing. Two recent eruptions in Chile, at Chaitén Volcano in 2008–10 and Cordón Caulle in 2011–12, allowed the first detailed observations of rhyolitic activity and provided previously hidden insights into the evolution of highly silicic eruptions. Both events exhibited simultaneous explosive and effusive activity, with both lava and ash plumes emitted from the same vent (Castro et al, 2014). The permeability of fracture networks that act as fluid flow pathways is key to such eruptive behaviour. We will investigate permeability systematically, at magmatic temperatures and pressures, in the presence of pore fluids using our newly-developed experimental capability.

ReferencesCastro, J M, Bindeman, I N, Tuffen, H and Schipper, I. (2014) EPSL 405, 52–61.

Gaunt, H E, Sammonds, P, Meredith, P G, Kilburn, C R J and Smith, R. (2013) IAVCEI 2013 Scientific Conference (1W_2K-P6), Kagoshima, Japan.

Rocchi, V, Sammonds P R and Kilburn C. (2004) J Volcanology Geotherml Research 132,137–157


66


HYDRAULIC FRACTURING — SIMULATION IN THE LABORATORY

St. Gehne1, P Benson1, N Koor1 and M Enfield2

1School of Earth and Environmental Science, University of Portsmouth, UK2PDF Ltd.

Significant unconventional hydrocarbon resources have been found in the UK (Andrews, 2013), which lead to a surge of interest in mudrocks and exploration techniques in recent years. Hydraulic fracturing is the key process to extract these hydrocarbon resources from the low permeability rock formations (e.g. Vinciguerra et al., 2004).

This study investigates the fracture mechanics behaviour of the fluid driven mechanical fracture process with a focus on the competition between permeability and overpressure and the derived fracture pattern. The key aims of the study are to establish the level of fluid overpressure needed to fracture the sample in tension, as well as the effect of the flow rate, with reference to the fluid that is carried away from the fracture zone by the natural rock permeability and to establish the conditions to generate a large network of small cracks, which is desired for optimum gas extraction.

A new technique has been developed, where modifications to the hydrofracture setup used by Vinciguerra et al. (2004) have been made to allow a fluid-rock contact and to generate micro tensile fractures from the inside of the sample. A suite of different porosity/permeability rock types such as Darley Dale sandstone (14% porosity) and Bentheim sandstone (22%) have been used for the initial experiments to prove the concept. The second phase extended the investigation from the relatively homogeneous rocks to the tight and highly anisotropic Crab Orchard Sandstone, to test the importance of inherent and induced anisotropy on the fracture pattern that evolved. Different stress environments have been applied to modify the inherent anisotropy so as to assess whether these environmental parameters influence geomechanical properties that are of importance for the ‘frackability’ of the rock and the fracture pattern.

Acoustic Emission (AE) location, successfully applied for similar studies (e.g. Benson et al., 2007 and Vinciguerra et al., 2004), is the key method used to accurately record and map the nucleation and development of the micro-fracture network.

Here we report for the first time, all of these data measured in the laboratory in well-controlled conditions and with new and innovative triaxial and AE equipment.

Further research is planned to use UK shale lithologies and to replicate the stress and temperature conditions found in the field, with reference to key structural and mineralogical rock features, and the structural geology and stress environments of potential source basins in the UK.


67


FRICTION OF PRINCIPAL SLIP SURFACES IN LIMESTONE: PRELIMINARY DATA FROM LARGE-SCALE BIAXIAL EXPERIMENTS

T Tesei1, B M Carpenter1, A Sagy2 and C Collettini3

1Istituto Nazionale di Geofisica e Vulcanologia, Rome, Italy2Geological Survey of Israel, Jerusalem, Israel

3Dipartimento di Scienze della Terra, Università La Sapienza, Rome, Italy

Principal slip zones (PSS) are localized shear surfaces that accommodate a large amount of displacement in brittle fault zones. Natural PSS are also characterized by several phenomena, at different scales, including fault plane grooves and comb and ring fractures, and commonly host extremely comminuted granular material. Moreover, recent investigations of PSS in carbonates (e.g. Fondriest et al., 2013, Siman-Tov et al., 2013, Collettini et al., 2013) suggested that PSS are the loci of seismic slip localization that induce peculiar microstructures such as mirror-like surfaces and thermal decomposition features caused by frictional heating.

To further investigate the frictional behaviour of PSS, we performed some preliminary experiments on large (20 cm x 20 cm) surfaces using a single direct shear configuration in the biaxial apparatus 'BRAVA' (INGV HP-HT lab, Rome). Experiments were carried out on slabs of limestone with dramatically different starting surface roughness (mean asperity height from 6 to 500 µm) in order to simulate the naturally-occurring roughness of carbonate-bearing PSS.

Mechanical data show that initially smooth surfaces are significantly weaker (µ~0.4) than rough surfaces (µ~0.9) and promote stick-slip of the sliding surface. Stick-slip behaviour is suppressed when wear material is accumulated between the surfaces and friction is greatly increased. Post-experimental observations document the heterogeneous development of a gamut of surface phenomena commonly recognized in natural fault surfaces such as fine-grained gouge organized in Riedel and boundary shears, oblate grooves decorated with ring fractures connected to Riedel shears.


68


INFLUENCE OF FRACTURE FILL ON PERMEABILITY IN MACRO-FRACTURED BASALT

Guangzeng Wang1,2, Thomas Mitchell2, Philip Meredith2, Pamela Perez-Flores3,4 and Yoshitaka Nara5

1School of Geosciences, China University of Petroleum (East China), Qingdao, China. 2Department of Earth Sciences, University College London, London, UK.

3Pontificia Universidad Católica de Chile, Santiago, Chile. 4Andean Geothermal Centre of Excellence (CEGA), Santiago, Chile.

5Graduate School of Engineering, Tottori University, Tottori, Japan.

Fractures are ubiquitous on all scales in crustal rocks and are commonly filled with detrital material; such as the comminuted gouge observed in many shear faults. Fracture networks also allow crystalline rocks to store and transport fluids. In a study on experimentally fractured basalt, Nara et al. (2011, Tectonophysics) showed that macro-fractures dominated permeability at low pressure and shallow depth, but that micro-fractures, which were harder to close, played an increasing role as pressure and depth increased. Their study, however, was limited to permeability in pristine, mated and unfilled fractures. We have therefore extended that work to study the effect of fracture-fill on permeability through macro-fractures in samples of the same material; Seljadur basalt, a fresh, columnar-jointed, intrusive basalt from SW Iceland with no visible pre-existing cracks and exceptionally low initial permeability.

We made steady-state flow permeability measurements on 38 mm diameter cores of Seljadur basalt which had been split in half using the Brazil disk technique to produce through-going axial macro-fractures across the diameter of previously intact material. Measurements were made under a range of different conditions: (1) Baseline measurements on unfilled but mated macro-fractures at effective pressures up to 60 MPa. (2) To explore the influence of fill material, measurements on the same macro-fractured sample but with an artificial fault gouge prepared from ground and sieved basalt particles with a maximum grain size of 60 µm. Measurements were made over the same pressure range with gouge thickness of 0.25 mm, 0.5 mm, 1.0 mm and 1.5 mm. (3) To investigate the influence of gouge grain size, measurements on the same sample with a 0.5mm thick gouge layer, but with a maximum grain size of 125 µm.

Similar to Nara et al. (2011), we find that the introduction of a through-going macro-fracture increases permeability by several orders of magnitude, but the permeability of the mated fracture decreases rapidly with increasing effective pressure. By contrast, the initial permeability of gouge-filled fractures was one to two orders of magnitude lower than unfilled fractures, and only decreased marginally as effective pressure was increased. The filled fracture with the thinnest gouge layer (0.25 mm) had the highest permeability, and once the gouge thickness exceeded 0.5 mm the permeability changed only slightly as a function of thickness. The grain size of the fill material also has an influence, with the permeability of the fracture containing gouge with a maximum grain size of 125 µm being significantly higher at every pressure than the same fracture with 60 µm maximum grain size gouge. Further, and unlike flow in unfilled fractures, we observed that, following the introduction of gouge, the permeability exhibited a directional dependence with fluid flow being significantly different for opposite flow directions.

Overall, the presence of a gouge-like layer of fill in macro-fractures decreases their permeability by several orders of magnitude relative to unfilled fractures. By contrast, the permeability of filled fractures only decreases marginally with increasing effective pressure, implying that compaction is only relatively minor and that the gouge continues to prop the crack open at elevated pressure. Permeability increases with fill thickness up to some critical value (0.5 mm in our case), but barely changes for thicknesses above that. We suggest that the critical gouge thickness is related to the fracture roughness; once the gouge layer is thick enough to control the flow, the rough fracture surfaces cease to exert any controlling influence. The observation of directional dependence of flow in gouge-filled fractures also suggests that gouge distribution and order also influences permeability significantly. In fact, post-mortem examination of some samples showed that fluid migration channels had developed in the gouge during flow. Not only will such features lead to enhanced permeability, but they will also lead to the development of a permeability anisotropy within the fracture.


69


TELESEISMIC HYDRO-SEISMOGRAMS AND SUSTAINED WATER LEVEL CHANGE

V Lyakhovsky1, I Kurzon1, M –L Doan2 and E Shalev1

1Geological Survey of Israel, 30 Malkhe Israel Street, Jerusalem 95501, Israel2Institut des Sciences de la Terre (ISTerre), University Joseph Fourier – Grenoble, France

Travelling seismic waves and Earth tides are known to cause oscillations in well water levels and in some occasions also sustained water level change. We describe and explain the behavior of two water wells: Gomè 1 well and Meizar 1 well, located along the active plate boundary, Dead Sea Transform. The water level in these wells oscillates in response to volumetric strain (P and Rayleigh waves) as well as to deviatoric strain (S- and Love-waves). In both wells, there is a sustained water level change to the distant September 24 2013 Mw=7.7 Pakistan (D= 26.469°) earthquake. In Gomè 1 well there is a ~600 Pa sustained pressure decrease while in Meizar 1 well there is a ~200 Pa sustained pressure increase. The water pressure oscillation is explained by the non-linear elastic behavior of the highly damaged rocks. Rock damage is calculated based on the coupling between water pressure and deviatoric strain response to surface waves. High water pressure oscillations responses to deviatoric strain at Gomè 1 and at Meizar 1 wells suggest that the rocks are heavily damaged. Visco-Poro-Elastic Damage Model for Brittle-Ductile Failure of Porous Rocks addresses two competing processes that may change the water pressure during deformation. Dilatency due to damage increase will decrease the water pressure whereas shear enhanced compaction will increase the water pressure. The pace of each process depends on material properties and on the damage state of the rock. In Gomè 1 well we observe sustained water pressure decrease due to higher pace of the damage induced dilation of sandstone. In Meizar 1 well compaction dominates and sustained water pressure increases due to higher pace of compaction.


70


71

POSTERS SESSION 2


72


THE ROLES OF COMPOSITIONAL HETEROGENEITY AND RESIDUAL MELT REMOBILIZATION IN STRAIN LOCALIZATION PROCESSES WITHIN

THE RUM VOLCANO LAYERED CUMULATE INTRUSION

A F Bell1 and N S Roberts1

1School of Geosciences, University of Edinburgh, UK

The rocks of the Isle of Rum, northwest Scotland, are famous for recording the history of the Paleogene Rum volcano. Though extensively studied for more than a hundred years, the geological variety and excellent exposures are still providing new insights into a wide range of volcanic and tectonic processes. The Rum layered intrusion consists primarily of a sequence of ultramafic and mafic cumulate rocks formed in an evolving magma chamber at depth below a central volcano. The Central Series (CS) is the youngest component of the layered intrusion, and is considerably more structurally complex than the relatively un-deformed sequences of cumulate rocks in the Eastern and Western Series. The rocks of the CS are dominated by a layered series of peridotites, gabbros, and anorthosites corresponding, at a broad scale, to the cumulate products of fractionation from a picritic magma. However, these rocks have been extensively deformed during and shortly after deposition by a combination of gravitational slumping, graben formation, and movements of the Long Loch Fault Zone which cuts the CS. Here we describe a variety of deformation features and textures within the CS, and identify different mechanisms that promote strain localization in variably consolidated cumulate rocks. Compositional and textural heterogeneity between, and within, units promotes strain localization in to specific weak horizons, particularly the gabbros and anorthosites. This deformation remobilizes feldspar-rich melts, which move upwards through the sequence along vertical strike-slip shear-zones, weakening these and promoting further movement. Elsewhere, where the deformation fabric is dipping significantly more steeply than the layering, high-angle normal faults develop, often entraining feldspar rich material into the deformation zone. High temperature gradients mean that both brittle and ductile processes and fabrics occur in close spatial and temporal proximity. These new observations help explain the wide range and variability of textures in the central series, and suggest that deformation can drive significant post-deposition evolution of bulk and trace-element compositions.


73


CRYSTAL PLASTICITY IN MAGMAS: AN EXPERIMENTAL STUDY

J E Kendrick1, E Mariani1, Y Lavallée1 and D B Dingwell2

1Department of Earth, Ocean and Ecological Sciences, University of Liverpool, UK2Department of Environmental Sciences, Ludwig Maxximilian University of Munich, Germany

Understanding the rheological behavior of magma during ascent and its influence on eruption style requires models, which themselves rely upon experimentally derived constraints. Models of suspension rheology have long attempted to deal with crystal fraction, shape and aspect ratio as rheological variables; however, recent advances in experimental magma deformation and imaging now provide a substantial opportunity for completing our picture of the viscous behaviour of multi-phase systems. This study reports the first observation of crystal plasticity, identified using electron backscatter diffraction (EBSD), in the plagioclase phenocrysts and microlites of natural andesitic magma from Volc án de Colima (Mexico). The same magmas then deformed experimentally at magmatic temperature (945°C) at constant stresses (12 or 24 MPa) to a total strain of up to 30% yield a further plastic response in the crystalline fraction, observable as a lattice misorientation, which grows with increasing stress and strain. Phenocrysts which contain brittle fractures show the highest values of crystal-plastic deformation in the intact segments. Evidently, crystal plasticity plays a role in strain accommodation under volcanically relevant conditions, bridging the viscous-brittle transition and leading crystal-bearing magmas towards failure. This behavior in crystal-rich ascending magma could favour strain localisation, shear zone formation and plug flow in the upper conduit.

Fig. 1. Misorientation, a measure of crystal-plastic deformation, across two plagioclase crystals(both 60 mm long): (left) from the as-collected magma from Volcán de Colima, and (right) from the same starting magma deformed at 945°C at 24 MPa to

30% strain, showing higher value of misorientation across the crystal.


74


A NEW CONCEPTUAL MODEL OF COMPACTION CREEP IN CARBONATE ROCKS AND APPLICATION TO A COMPACTING RESERVOIR

D Keszthelyi1, B Jamtveit1 and D K Dysthe1

1Physics of Geological Processes, Department of Physics and Department of Geosciences, University of Oslo, Norway

Rocks subject to compressional or shear stresses can deform slowly and irreversibly during time. In large scale this can be observed as compacting reservoirs due to fluid (hydrocarbon or water) production and creeping faults at strike-slip plate boundaries.

We created a simple conceptual micromechanical model of compaction creep in rocks under hydrostatic conditions. This model combines microscopic fracturing and pressure solution and if scaled up to macroscopic scale by a statistical approach it can be used to predict strain rate at core scale. The model uses no fitting parameter and has few input parameters: effective stress, porosity, pore size distribution, temperature and water saturation. Internal parameters are Young’s modulus, interfacial energy of wet calcite and the dissolution, diffusion and precipitation rates of calcite, all of which are measurable independently. The model was tested against existing long-term creep experiments and it was able to predict the magnitude of the resulting strain under largely different effective stress, temperature and water saturation conditions.

The model was also able to predict the observed compaction of the Ekofisk field an example of a chalk reservoir in the North Sea: oil production led to a significant pore pressure drop (effective stress increase) inside the reservoir causing compaction and the subsequent subsidence of the seafloor. Using pressure history data subsidence rate was calculated for the centre of the field and the calculations showed a good agreement with the field observations.

We believe that further generalization of the model might function as a general theory of long-term creep of rocks in compressional settings and can be used to predict subsidence at other producing fields with different geological settings.

Figure 1: A schematical view of the Ekofisk field subsidence, the proposed model and the comparison of predicted and observed subsidence data


75


SITE CHARACTERIZATION OF SELECTED ACCELEROMETRIC STATIONS’ SITES IN CRETE ISLAND (GREECE) FOUNDED ON ROCK AND

SOFT ROCK FORMATIONS

C Loupasakis1, P Tsagaratos1, D Rozos1, A Vafidis2, M Steiakakis2, A Savvaidis3, P Soupios4, I Papadopoulos4, N Papadopoulos5, A Sarris5, M-D Mangriotis6 and U Dikmen7

1School of Mining and Metallurgical Engineering, National Technical University of Athens, Greece;2Department of Mineral Resources Engineering, Technical University of Crete, Greece

3Institute of Engineering Seismology and Earthquake Engineering (EPPO), Thessaloniki, Greece4Department of Natural Resources and Environment, Technological Educational Institute of Crete, Chania, Grete, Greece

5Laboratory of Geophysical-Remote Sensing & Archaeoenvironment, Institute for Mediterranean Studies, Foundation for Research & Technology Hellas, Greece

6Institute of Petroleum Engineering, Heriot-Watt University, UK7Department of Geophysics, Ankara University, Turkey

The design of civil constructions requires a detailed knowledge of the near surface ground conditions. The bearing capacity and the stress — strain relations as well as the ground amplification and corresponding peak ground motion can be determined by investigating the static and dynamic geotechnical parameters as well as the ground type category. This knowledge can be obtained by combining geotechnical and geophysical methods, such as engineering geological surface mapping, geotechnical drilling, in situ and laboratory testing and geophysical investigations.

The above mentioned methods were combined for the site characterization in selected sites of the Hellenic Accelerometric Network (HAN) in the area of Crete Island. The combination of the geotechnical and geophysical methods in thirteen (13) sites provided sufficient information allowing the accurate site characterization. On the other hand the combination of the methods determined their limitations, setting up the minimum tests requirements in relation to the type of the geological formations. The investigation was mainly focused on the characterization of sites located on soft rock Neogene formations as well as on Mesozoic rock formations, while some sites were also located on soil formations.

The reduced accuracy of the surface mapping in urban sites, the uncertainties introduced by the geophysical survey in sites with complex geology and the 1D data provided by the geotechnical drills are some of the causes affecting the efficiency of the investigation methods. Through this study the gradual improvement on the accuracy of the site characterization data in regards to the applied investigation techniques is presented by providing characteristic examples from the total number of thirteen sites.

This research has been co-financed by the European Union (European Social Fund — ESF) and Greek national funds through the Operational Program 'Education and Lifelong Learning' of the National Strategic Reference Framework (NSRF) — Research Funding Program: THALES. Investing in knowledge society through the European Social Fund.


76


ADVECTION-DISPERSION OF A PASSIVE TRACER IN NETWORKS OF PIPES: EFFECT OF CONNECTIVITY

Y Bernabé1*, Y Wang2, M Li2

1Earth, Atmospheric and Planetary Sciences Department, MIT, Cambridge, USA2State Key Laboratory of Oil and Gas Reservoir Geology and Exploitation, Southwest Petroleum University, Chengdu, China

The purpose of this work was to model advection/dispersion at the macroscale as a result of the mechanisms acting at the microscale (i.e., more precisely, at the pore scale). We simulated fluid flow through SC, BCC and FCC pipe network realizations with different pipe radius distributions and different levels of connectivity as measured by the pore coordination number z (see Bernabé et al., 2011). We used periodic boundary conditions, so the advective transport of a passive tracer could be simulated over a very large periodic array of identical network realizations. In order to simulate dispersion, we assumed that individual tracer particles obeyed Taylor dispersion. When a particle enters a pipe, we randomly assign it a residence time satisfying the local Taylor dispersion probability function (Van Genuchten and Alves, 1982). When the particle exits, it randomly proceeds into one of the pipes connected to the original one according to probabilities proportional to the outgoing volumetric flow in each pipe. We thus simulated the advective/dispersive motion of up to 10000 particles in different periodic arrays of network realizations with specific values of pore connectivity and pore size heterogeneity. The mean velocity of the particles was 0.001 ms-1 and the distance travelled was on the order of 10 m. The longitudinal and transverse dispersion coefficients were calculated by the method of moments. One important observation was that SC, BCC and FCC networks, constructed to have identical connectivities and pore size heterogeneity levels, displayed significantly different dispersion coefficients. Hence, unlike permeability or formation factor (Bernabé et al., 2011), dispersion is not a 'universal' property of pipe networks. However, a number of general observations can still be made. Longitudinal dispersion is at least one order of magnitude greater than transverse dispersion and both strongly increase with decreasing pore connectivity. In conditions of fixed pore connectivity and pore size heterogeneity, the dispersion coefficients increase as power laws of the mean pipe radius, or, in other words, of the network permeability, in agreement with the experimental data of Harleman et al. (1963).

ReferencesBernabé, Y, M Zamora, M Li, A Maineult and Y B Tang (2011), Pore connectivity, permeability and electrical formation factor: a new model and comparison to experimental data, J. Geophys. Res., 116, B11204, doi:10.1029/2011JB008543.

Harleman, D R F, Melhorn P F and Rumer R R. (1963), Dispersion-permeability correlation in porous media, J. Hydr. Div., Proc. ASCE, 89(HY2), 67–85.

Van Genuchten, M T and W J Alves (1982), Analytical solutions of the one-dimensional convective-dispersive solute transport equation, US Dept. Agriculture Tech. Bull., 1661, 156 pp.


77


THE EFFECT OF CHEMICAL ENVIRONMENT ON THE TIME-DEPENDENT COMPACTION BEHAVIOUR OF QUARTZ SANDS

M T W Schimmel1,2, S J T Hangx1 and C J Spiers1

1HPT Laboratory, Department of Earth Sciences, Faculty of Geosciences, Utrecht University, Netherlands2Energy and Resources, Copernicus Institute for Sustainable Development, Faculty of Geosciences,

Utrecht University, Netherlands

Hydrocarbons play an important role in the energy system and will remain doing so for some time in the future. However, in some locations, decades of hydrocarbon production have resulted in reservoir compaction and surface subsidence. Such effects can have significant unwanted impact on the environment and surface infrastructure. Aside from poro-elastic effects, which are relatively easily predictable, compaction of reservoir rocks can potentially also be time-dependent due to the time-dependent deformation (creep) of the rock matrix. Previous studies on compaction of both clastic and carbonate reservoir rocks and model aggregates have shown that the chemical environment plays an important role in determining the extent of creep. The aim of this research is to further assess the effect of chemical environment on creep in quartz aggregates.

To do this, we are conducting uniaxial compaction experiments on aggregates of Ottawa quartz sand (d = 212–250 µm), at room temperature conditions or at elevated temperature of 75°C to mimic reservoir conditions. The experiments consist of two stages: 1) load cycling up to 35 MPa effective stress under lab dry conditions, and 2) creep at a constant effective stress of 35 MPa. During the creep stage, the sample is either dry (lab air, dry CO2), or vacuum-flooded (silica-saturated solution, wet CO2, 0.1 M aluminium chloride solution). The pore fluids and gases (except for lab air) are introduced and maintained during the experiment at 10 MPa total pressure.

The experiments show that wet experiments produce more creep (lab air vs silica-saturated solution), while less creep occurs in the CO2 and aluminium chloride experiments compared to the silica-saturated solution experiment. The observations suggest that the time-dependent deformation of the quartz aggregate is mainly related to subcritical crack growth or stress corrosion cracking, resulting in grain failure and subsequent rearrangement of grain fragments. The inhibiting effect of CO2 is inferred to be related to the acidic environment CO2 creates, which reduces the crack growth rate. A similar effect is inferred to occur in the presence of aluminium chloride. Apart from creating an acidic environment, it has been suggested that aluminium ions reduce the solubility of quartz, which further reduces the rate at which subcritical cracking can occur. The results suggest that perhaps acidic additives or CO2 injection may have applications in reducing creep in clastic reservoirs.


78


GEOMECHANICS OF FRACTURE AND KAISER EFFECT

Sunjay1 and Manas Banerjee1

1Department of Geophysics, Banaras Hindu University, India

Kaiser effect is one of the most important and interesting manifestations of the fundamental ability of rocks and materials to accumulate, to retain and to reproduce information on the peak stresses and strains experienced in the past. This ability called ‘stress memory’ or ‘endpoint-memory’ has attracted attention as a possible basis for stress measurement in rocks. Acoustic and electromagnetic emissions in rocks is associated with micro-cracks. Rocks break in either tension, resulting in tensile fractures, or compression, resulting in shear fractures. Fracture geometry is usually described by fracture height, length and width (or aperture). Fracture propagation is an increase in fracture length and height. An increase in fracture aperture due to increased fluid pressure (natural or as part of operations such as hydraulic fracturing) is called dilation. In situ stress states are usually divided into three types based on the relative magnitudes of the three principal stresses. If the vertical stress is the maximum stress, the regime is normal faulting. If the vertical stress is the intermediate stress, the regime is strike-slip faulting. If the vertical stress is the least stress, then the regime is reverse faulting. Tectonic fractures, related to folding and faulting, tend to be variably oriented and regionally inconsistent and may be related to a past in situ stress environment, not the current day stress state. If we know the orientation and magnitudes of the in situ stresses, we use the orientation of the fracture or fault plane to calculate the shear and normal stresses acting on it. We plot a Mohr diagram to describe slip on fractures and faults. Visco-elastics rock(salt) , visco-aniso-elastic properties study is challenging task before geoscientist. Wavelet analysis is employed for investigation of rock physics and geomechanics of microseismic , hydraulic fracturing and fracture propagation in hydrocarbon reservoir. Coal cleat fracture propagation study is very important for coal bed methane explotation. G-function analysis is employed for fracture propagation.


79


MEASURING STRESS IN ROCK USING ELECTRIC POTENTIAL SENSOR TECHNOLOGY

J W Archer1, M R Dobbs2, H J Reeves2 and R J Prance1

1Sensor Technology Research Centre, Department of Engineering and Design, University of Sussex, Brighton, Sussex, BN1 9RH, UK2British Geological Survey, Environmental Science Centre, Keyworth, Nottingham, NG12 5GG, UK

Knowledge of in-situ stress state is critical for predicting the response of rock masses when subjected to natural and man-made disturbances. It is therefore of direct relevance to civil, mining and petroleum engineering, and also to geology and geophysics in general [1–3]. In situ stress may be estimated theoretically using constitutive equations or may be estimated indirectly by disturbing the rock mass. Current methods used include jacking, hydraulic fracturing, relief methods, strain recovery methods, and borehole breakout methods [1–4]. Both theoretical and empirical methods have a number of drawbacks including: dependence on assumptions of elasticity, homogeneity and isotropy [4]; accuracy at best of ± 20% [1]; high cost due to use of specialised equipment; invasive procedures that disturb the ground [4]; logistical constraints associated with instrumentation relative to principal stress; and limitation to single instantaneous measurements.

One technology that could resolve many of these issues is the Electric Potential Sensor (EPS) developed at the University of Sussex (UoS)[5]. Previously, laboratory-based studies undertaken by UoS and the British Geological Survey (BGS) have utilised EPS to measure pressure stimulated voltages (PSV’s) of different lithology types up to and during failure [6–7]. The hypothesis of this work is supported by laboratory-based studies that have found the flow of electric current in rock is associated with deformation and the application of stress [8–12]. These PSV studies used capacitively coupled EPS that are only capable of measuring the AC components of PSV’s, meaning only changes in PSV could be measured.

In order to measure ‘static’ stress UoS have developed resistively coupled EPS technology that is capable of measuring the DC component of PSV’s. UoS and BGS have already undertaken preliminary experimental testing to investigate the relationship between PSV and applied pressure. The experimental method involved subjecting a 30 mm cubic specimen of marble to progressive step-like uniaxial loading using a pneumatic actuator mounted in a small load frame. The resulting PSVs were recorded using directly coupled EPS. For each stress step one data point was selected where dv/dt ≈ 0. The selected data points were used for polynomial regression analysis between applied pressure and PSV.

The resulting R2 and adjusted R2 values are 0.9279 and 0.9119 respectively. It is evident that the R2 of the regression is high, thus demonstrating that resistively coupled EPS measurement of PSV’s can be used to monitor stress in specimen of rock subject to uniaxial loading.

The proof of principle uniaxial loading experiments demonstrate the possible application of EPS technology for measuring in-situ stress. Furthermore, due to the technology’s low cost, amenity to deployment in sensor arrays, and non-invasive nature [10], it has the potential to significantly improve upon traditional methods of in-situ stress measurement.]

ReferencesAmadei, B and Stephansson, O. 1997. Rock Stress and its Measurement. Chapman & Hall, London and New York.

Fairhurst, C. 2003. Stress estimation in rock: a brief history and review, International Journal of Rock Mechanics & Mining Sciences, 40, 957–973.

Jaeger, J C, Cook, N G W and Zimmerman, R W. 2007. Fundamentals of Rock Mechanics (4th ed.). Blackwell Publishing, Oxford.

Farmer, I W. 1983. Engineering Behaviour of Rocks (2nd ed.). Chapman and hall, London.


80

Prance, R J, A Debray, T D Clark, H Prance, M Nock, C J Harland and A J Clippingdale. 2000. An ultra-low-noise electrical-potential probe for human-body scanning. Measurement Science & Technology, 11: 291–297.

Aydin, A, R J Prance, H Prance and C J Harland. 2009. Observation of pressure stimulated voltages in rocks using an electric potential sensor. Applied Physics Letters, 95: 124102.

Aydin, A, Dobbs, M R, Reeves, H J, Kirkham, M P and Graham, C C. 2013. Stress induced electric field measurements of different rock lithology using the Electric Potential Sensor. 47th US Rock Mechanics / Geomechanics Symposium. 23–26th June 2013. San Francisco, USA.

Triantis, D, I Stavrakas, C Anastasiadis, A Kyriazopoulos and F Vallianatos. 2006. An analysis of pressure stimulated currents (PSC), in marble samples under mechanical stress. Physics and Chemistry of the Earth, 31: 234–239.

Hadjicontis, V., C. Mavromatou, and D. Ninos. 2004. Stress induced polarization currents and electromagnetic emission from rocks and ionic crystals, accompanying their deformation. Natural Hazards and Earth System Sciences, 633–639.

W Gebrial, R J Prance, C J Harland and T D Clark. Noninvasive imaging using an array of electric potential sensors. REVIEW OF SCIENTIFIC INSTRUMENTS 77, 063708 (2006).


81


CONTAIN: THE GEOMECHANICS OF CANDIDATE CCS RESERVOIRS AND SEALS

M R Dobbs, A Ougier-Simonin, C C Graham, R J Cuss and J F Harrington

British Geological Survey, Keyworth, NG12 5GG, UK

Carbon Capture and Storage (CCS) is proposed as a key technology for limiting CO2 emissions to tackle climate change, while potentially offering economic benefits. Depleted hydrocarbon reservoirs represent a significant resource for storage of CO2 in offshore UK. The research work combines an experimental numerical approaches in order to examine the impacts of depletion and injection on both the reservoir and the overlying succession of sealing rocks. A social science research component is also included, investigating the effectiveness of public understanding and engagement methodologies.

As part of the project’s experimental programme, a series of rock deformation tests have been conducted to investigate the geomechanical and geophysical properties of samples from the Sherwood Sandstone Group (10–18% porosity). We present the mechanical results of a suite of uniaxial, triaxial and hydrostatic deformation tests, at effective pressures from 0–160 MPa, and axial deformation strain rates from10-5 -10-6 s-1. The experimental set-up enables us to monitor and independently vary pressures, axial stress and temperature (0–180°C). Hydraulic permeabiltiy measurements, as well as P and S elastic wave velocities can also be monitored during testing. The resulting data-set allows delineation of the critical state failure envelope for Sherwood Sandstone, based on estimated yield stresses. This work complements findings from additional experiments conducted on samples taken from the Mercia Mudstone Group, as well as numerical simulation activities at laboratory to field scale. Future work will involve the use of critical state data to conduct scenario analysis for a range of stress paths simulating depletion and injection activities. The results will also inform numerical modelling for storage site performance assessment. CONTAIN is an EPSRC funded project.


82


IMPROVED IDENTIFICATION OF CORRELATIONS IN NATURALLY VARIABLE MATERIALS BY MULTISTEP MECHANICAL TESTS

J-M Hertzsch, W Gräsle and A Kaufhold

Federal Institute for Geosciences and Natural Resources, Stilleweg 2, 30655 Hannover, Germany

Various mechanical properties (e.g. strength parameters, elastic parameters) of rocks cannot be characterized by constant parameters, but require a description as functions of state variables. Usually, testing a large number of samples is necessary to identify the functional relationship and its parameters for every single material. The natural variability of rock is a common disadvantage of this approach, since it blurs the functional relationship and may increase the number of required tests beyond practicability.

Obtaining multiple data sets from a single rock sample would allow to circumvent this problem, in particular if the identification of the functional dependency of a mechanical material property on the state variables is required. Unfortunately, changes of material properties due to progressive damage of the sample in many types of mechanical testing (e.g. standard methods for determination of strength parameters) severely limit the amount of data available to characterise the undamaged material.

We present two examples of improved testing routines developed to overcome this problem: The investigation of Young’s modulus E and of shear strength peak, both as functions of the confining pressure 3. The methods are distinguished by the effort to avoid material damage during the test or at least to limit it to an acceptable degree. Any presented investigations were carried out on large samples (Ø100 mm) of Opalinus Clay from Mont Terri Underground Rock Laboratory (Switzerland).

Figure 1: The dependency of Young’s modulus from confining pressure for Mont Terri Opalinus Clay. Whereas loading normal

to the bedding (or at an intermediate inclination of 55°) results in a linear-relationship, samples loaded parallel to the bedding display a significantly nonlinear behaviour sufficiently described by a power law.


83

Avoiding sample damage during the determination of Young’s modulus under triaxial conditions (in a Kármán cell) is a mandatory requirement in the standard method and can be accomplished quite easily. Nevertheless, for claystone usually only one value of E per sample has been determined, and the E(3) relationship has been largely neglected so far. Our multistep tests revealed a significant dependency of E on 3 for Opalinus Clay. It could be demonstrated that the anisotropy of the claystone does not only affect the parameters of Young’s modulus but also the functional characteristic of the E(3)-relationship. While samples loaded perpendicular to the bedding planes or at an intermediate inclination display a linear relationship E(3)= E0+ 3, results from loading parallel to the bedding are best represented by a power law E(3)=c 3

b (Fig. 1).

Investigating the relationship of the shear strength peak on the confining pressure 3 on a single sample is a much more difficult task. Obviously, damaging the sample during the determination of shear strength is unavoidable. Thus, sample degradation can only be limited by unloading immediately after detecting the failure condition. A routine has been developed to check whether results from repeated testing of the failure condition are sufficiently representative for the undamaged material. We found that up to 4 consecutive measurements can be considered valid for samples from the shaly facies of Mont Terri Opalinus Clay. However, repeated testing was not successful for material from the sandy facies of Opalinus Clay due to its more brittle behaviour. For the shaly facies a linear branch of the peak (3)-relationship at low confining pressure (typically 3 <7.5 MPa) can be identified. At higher confining pressure the shear strength falls below the values predicted by this linear relationship, probably constituting a second linear branch with significantly lower slope (Fig. 2).

Figure 2: Up to 4 subsequent valid measurements of failure strength could be carried out on a single sample of Mont Terri Opalinus Clay (shaly facies). An overall nonlinear-relationship including an initial linear branch

and a second probably linear branch was found.


84


EXPERIMENTAL LIMITATIONS ON THE MEASUREMENT OF ELASTIC DISPERSION: MODELLING THE EXPERIMENT

L Pimienta1, J Borgomano1, J Fortin1 and Y Guéguen1

1laboratoire de Géologie de l’ENS - PSL Research University - UMR8538 du CNRS.

The dispersion/attenuation phenomena in sedimentary rocks occur from fluid flow at different scales. Different experimental procedures have been developed to investigate those phenomena, yet little is known on the possible boundary effects that need to be accounted for. The effects of the boundary conditions affecting the stress-strain experiment are here theoretically investigated.

Within the framework of linear/quasi-static poroelastic theory, using the equilibrium conditions and the mass conservation, it is possible to define the partial derivative equation satisfied by the pore fluid pressure. The source term used is a periodic 'isotropic solicitation' that corresponds to confining pressure oscillations. The model is analytically solved in 1D, in the direction of the sample’s length, leading to the volumetric strain that allows reaching both bulk modulus and attenuation. Two experimental conditions are tested: position of measurement gauge and effect of the dead volume.

Solving the equation for drained boundary conditions, the 1D model solution shows dependence of the calculated modulus on the position on the sample’s length. By integrating either over the sample’s total length (i.e. 80 mm) or over the strain gauge’s typical length (i.e. 6 mm), the difference between local and global measurement is tested. Furthermore, different positions for local measurement are also tested (Fig. 1). Predictions show a clear dependence to the measuring position and condition (i.e. local versus global). The transition for global predictions spans over 4 orders of magnitude in frequency. On the reverse, the local prediction at the sample’s centre resembles a Zener-like typical dispersion/attenuation mechanism.

Figure 1: Predicted bulk modulus (a) dispersion and (b) attenuation using the 1D model with drained boundary conditions.

The parameters used for the prediction are of a fluid-saturated sandstone of 7% porosity, with a drained bulk modulus of 14 GPa and permeability of 10–14 m2.

The effect of an existing dead volume at both ends of the sample is then tested through changing the boundary conditions. It is shown that, depending on the relative storage capacities of the sample and dead volume, the value measured under 'experimentally drained' conditions is in fact in between the purely drained and undrained values.


85


ELECTRIC CURRENT FLOW IN CARRARA MARBLE DURING TRIAXIAL DEFORMATION

A Cartwright-Taylor1, P Sammonds1 and F Vallianatos2

1Institute for Risk and Disaster Reduction, University College London, UK2Laboratory of Geophysics and Seismology, Technological Educational Institute of Crete, Greece

Understanding the generation of electric currents in the crust is key to understanding the wide range of observed pre-seismic geoelectric anomalies measured in the field. The underlying causes of these anomalies and their relationship with deformation are still largely unknown, and their statistical significance as earthquake precursors remains controversial. However, laboratory experiments have shown that piezoelectric and electrokinetic effects during rock deformation lead to both precursory and coseismic electric signals. Such signals have also been observed during deformation of dry, quartz-free rocks. In particular, spontaneously generated electric currents have been observed during uniaxial deformation of dry marble samples. Since marble is non-piezoelectric, the main question is: how are these electric currents related to deformation and damage? Here, we report results from a suite of triaxial compression experiments on Carrara marble samples. Differential electric current flow through the samples was measured during deformation at strain rates from 10-6 to 10-4 s-1, confining pressures of 10 to 100 MPa and with two types of pore fluid (pure water and ionic brine). Mechanical data, ultrasonic velocities and acoustic emissions were acquired simultaneously to constrain the relationship between electric current and damage. We examine how these currents evolve with deformation under crustal conditions and how they vary with deformation mechanism across the brittle-ductile transition.

Small (nano Ampere) electric currents are generated and sustained during deformation under all the conditions tested. Spontaneous electric current flow is due to the presence of localised electric dipoles. In the dry samples it is seen only in the region of permanent deformation and is correlated to the damage induced by microcracking, with a contribution from other intermittent ductile mechanisms. Onset occurs only after a 10% reduction in VP, implying a certain degree of crack damage and/or connectivity is required before current will flow. Three separate time-scales are apparent in its evolution and both absolute and fluctuating components of the signal are related to stress, damage, deformation mechanism and localisation of deformation leading to sample failure, and it exhibits a precursory change as the stress drop accelerates towards failure. Both current and charge production depend strongly on the experimental condition. Power-law relationships are seen with confining pressure and strain rate, with the first corresponding to microcrack suppression and activation of crystal plasticity, and the second to time-dependent differences in deformation mechanism, across the brittle-ductile transition. In the presence of an ionic pore fluid, electrokinetic effects dominate deformation processes, but development of the crack network drives the variation in electrokinetic parameters. In dry samples, current flow is approximately proportional to stress within 90% of peak stress. In fluid-saturated samples, proportionality holds from 40%, with a significant increase in the rate of current production from 90%, and is associated with fluid flow during dilatancy. This proportionality suggests that the electric signals could be used as a proxy for stress and, together with the power-law relationship between current and strain rate, is reminiscent of power-law creep.


86


A BIAXIAL PLANE STRAIN APPARATUS FOR NEUTRON DIFFRACTION-BASED EXPERIMENTS ON GRANULAR ROCKS

S D Athanasopoulos1, S A Hall1,2, A Nordin3, G Nikoleris3, G Couples4, J F Kelleher5 and T Pirling6

1Division of Solid Mechanics, Lund University, Lund, Sweden 2European Spallation Source AB, Lund, Sweden

3Division of Product Development, Lund University, Lund, Sweden 4Institute of Petroleum Engineering, Heriot-Watt University, Edinburgh, United Kingdom

5ISIS, Rutherford Appleton Laboratory, Chilton, Didcot, United Kingdom 6Institut Laue Langevin, Grenoble, France

This work considers the development of a new 'plane-strain' biaxial loading device for granular rocks through which the full-field investigation of strain evolution at different scales will be possible. Multiscale strain measurements will be accomplished by combining Neutron Diffraction with Digital Image Correlation (DIC) during plane-strain loading. The experimental set-up will also, in the future, include ultrasonic tomography to monitor the full-filed evolution of elastic properties.

Neutron Diffraction scanning has recently been successfully used to investigate force/stress distribution in granular materials under load. More specifically, Hall et al. (2011) showed that grain strains can be measured over a small gauge volume of a sample consisting of tens-of-thousands of sand grains. Further to that, Wensrich et al. (2012) produced in-situ mapping of the distribution of stress as an average over the volume of particles of a copper powder inside a solid die, excluding the voids. The key aim of this work is to extend the approach of Hall et al. (2011) to map spatial variations and evolutions of granular strains in rocks under loading, to investigate how forces are transmitted through the material and how this evolves with (localised) deformation. The simultaneous measurement of total strain fields (including porosity changes) through DIC and the use of samples with different cementations, will allow the characterisation of the mechanisms that act at different stages of deformation towards failure and how this is influenced by the degree of cementation.

Combining the different experimental techniques in a single apparatus requires certain constraints imposed by the different techniques and their combination to be addressed. A characteristic example relates to the combination of the design demands of the neutron measurements and of the high pressure needed to perform experiments under realistic in-situ conditions. The first requires the device walls to be as thin as possible to allow the maximum number of neutrons to reach the sample, whereas the second requires the walls to be thick enough to sustain the required confining pressures. A first prototype of the device, without high pressure, has been tested (including for neutron penetration) at UK’s neutron facility, ISIS. Results from this first proof-of-concept experiment, including 2D grain strain mappings for prismatic samples of sand loaded over a load-unload cycle, will be presented. Furthermore, the optimisation of the device, in terms of both mechanics and neutron scattering, for the construction of the second version, with confining pressure, will be presented and results of a first experiment with this device at the ILL neutron source in France will be discussed.

References Hall S A. et al., 2011, Granular Matter, 13, 251–254

Wensrich C M. et al., 2012, Granular Matter, 14, 671–680


87


FRICTIONAL AND MECHANICAL PROPERTIES OF VOLCANIC AND SEDIMENTARY ROCKS. APPLICATION TO MT ETNA (SICILY)

A Castagna1, S Vinciguerra1,2, N De Paola3, S Nielsen3, P Benson4, R J Walker1 and A Ougier-Simonin2

1Department of Geology, University of Leicester, UK2Rock&Soil Physics Laboratory, British Geological Survey, Keyworth, UK

3Department of Earth Science, University of Durham, UK4School of Earth and Environmental Science, University of Portsmouth, UK

Sliding processes and flank instability affect volcanic edifices worldwide. Mt. Etna, Sicily, is Europe’s tallest active volcano, and is subject to flank instability affecting its eastern sector. Sliding is confined in its northern-eastern part by the Pernicana Fault System (PFS). This fault system is the most active part of the entire flank with a slip rate of about 2 cm/y. Along its 20 km length, the PFS switch from a stick-slip (e.g. upper section of the system, near the summit) to an aseismic creep (e.g. lower part, toward the Ionian Sea) behaviour. The depth extent is unknown, but it is thought that the PFS cuts the entire volcanic pile, decreasing in dip into the sedimentary basement. The geometry and scale of displacement involves juxtaposition of the lithologies present, including the basaltic volcanic pile composing the edifice, the quaternary deposit of clay presents underneath the volcano, the sedimentary formations belonging to the Appenninic-Maghrebian Chain units (Europe domain) and the limestone belonging to the Hyblean plateau (Africa domain). Here, we report new mechanical tests (UCS, Multistage Triaxial Tests, Shear Box) to characterize strength, fracture nucleation and propagation, and the deformation mechanisms of the main lithologies listed above. In the future we aim to investigate the frictional properties of these lithologies. Pilot frictional data using rotary shear apparatus show promise: we present slide-hold-slide and velocity step tests on basalt-basalt and limestone-limestone at low velocities (13–130 µm/s) and normal stresses of 5 MPa and 20 MPa. The ultimate aim of this project is to explore the lithological control on the frictional behaviour of the Pernicana Fault System.


88


THE RESPONSE OF TUNNEL PERIMETER SUB-CRITICAL CRACK GROWTH TO OVERPRINTED STRESS AND THERMAL VARIATION

S Francis1 and P Benson1,2

1Department of Earth and Environmental Science, University of Portsmouth, UK2Director of Rock Mechanics Laboratory, University of Portsmouth, UK

A recent white paper published by the UK Government sets out a framework for the siting of a long term management facility for the geological disposal of high activity nuclear waste. Such a facility may be located at depths of up to 1 km and pose significant challenges during its construction, operation and closure phases due to geological stress conditions not commonly encountered in the UK civil engineering industry. The lifecycle of the disposal facility is such that it is required to function safely for a time period in the order of hundreds of thousands of years in a host rock containing pore water.

The influence of pore water as a chemically active fluid within a host rock promotes time dependant brittle deformation through sub-critical crack growth, the primary element of which is stress corrosion. Stress corrosion allows the deformation and subsequent failure of a rock at stresses considerably lower than its short term failure strength, and may be categorised in three phases: (i) Primary creep identified by decelerating strain (ii) Steady state creep and (iii) Tertiary creep characterised by accelerating strain due to crack interaction.

Here we report new data from a laboratory study where the effect of an overprinted stress on the inherent rock mechanical behaviour was simulated. This was achieved using a triaxial apparatus at elevated confining pressures (Pc = 27 MPa, approximately 1 km) and temperatures, and using a sample featuring an artificial cavity to represent an excavation. P-wave elastic velocities and Acoustic Emission (AE) were recorded continuously during the experiments, which were performed both in constant strain mode and in CREEP mode, with AE hypocenters located using a double difference methodology for additional accuracy (sum-mm). We found that the deformation speeds up significantly with increasing temperatures, and that AE is preferentially clustered in the top of the cavity. In an engineered nuclear waste repository, elevated temperatures are likely due to the radiogenic decay of high activity waste in addition to the geothermal gradient. Future investigations will focus on the required design element needed to arrest rock creep (time-to-failure) and extend the work to crystalline formations.


89


DAMAGE LOCALISATION AND VELOCITY EVOLUTION IN EL HIERRO AND TENERIFE BASALTS

C Harnett1, P Benson1 and M Fazio1

1Rock Mechanics Laboratory, School of Earth and Environmental Sciences, University of Portsmouth, Portsmouth, UK.

A large number of flank collapses are known, including those in El Hierro and Tenerife (Canary Islands). Collapses on volcanic islands are likely to be complex as they involve the combination of active tectonics, heat, and fluids. Not only does the volcanic system generate stresses that reach close to the failure strength of the rocks involved, but when combined with active pore fluid the process of stress corrosion allows the rock mass to deform and 'creep' at stresses far lower than the nominal short-term failure strength of rock. We present new data using rocks taken from El Hierro and Tenerife in order to better understand the evolution of acoustic emission velocities and the overall implication of this on flank instability. Experiments were conducted over short (30–60 minute) and long (8–10 hour) time scales. For this, we use the method of Heap et al. (2011) to impose a constant stress (creep) domain deformation monitored via non-contact axial displacement transducers.

Despite the obvious geological hazard posed by edifice failure, the phenomenon of creep in volcanic rocks has yet to be thoroughly investigated in a well-controlled laboratory setting. This is achieved via a conventional triaxial cell to impose shallow conditions of pressure (<25 MPa), and equipped with a 3D laboratory seismicity array (known as acoustic emission, AE) to monitor the micro cracking due to the imposed deformation. This examination of velocity change focusses on the time periods both before and after failure.

By equipping the sample assembly with an array of 12 piezo-electric Acoustic Emission (AE) sensors, we focus here on the generation of a geophysical image of the evolving damage as a function of strain. This is achieved by both active (velocity surveys) and passive (AE hypocentre location) means using a time varying anisotropic velocity model. We present new data illustrating that elastic velocity data are extremely sensitive to the changing strain evolution, and that velocities at high angles to the fault zone are less affected than raypaths propagating at a shallow incidence. We also discuss the post-failure recovery of the basalts and the impact of this on P-wave velocities, alongside an exploration of the timescale of velocity evolution prior to fault rupture. We link these observations to the damage and fracturing in the damage zone via thin section analysis.


90


A NEW METHODOLOGY FOR PALEOSTRESS RECONSTRUCTION USING THEORY, FIELD OBSERVATIONS AND PETROPHYSICAL DATA

Luca Smeraglia1, Fabio Trippetta1, Eugenio Carminati1,2, and Silvio Mollo3

1Department of Earth Science, 'Sapienza' University of Rome, Italy2National Research Council, Rome, Italy

3National Institute of Volcanology and Geophysics, INGV, Rome, Italy

The measurement of crustal stress magnitude is always challenging and generally poorly constrained. This is particularly significant in active fault zones where the knowledge of stress magnitude is crucial for understanding fault mechanics during earthquakes nucleation. In this work we propose a workflow using laboratory and field data as a proxy for quantitative paleostress reconstruction along active fault zone. We studied the exhumed Olevano-Antrodoco Thrust Fault (OATF) in Central Italy consisting of a SW-dipping thrust fault that juxtapose middle Miocene carbonates in the hangingwall above upper Miocene foredeep sandstones, W-SW-dipping, in the footwall. We collected 26 samples of footwall sandstones approaching progressively the OATF, from the undeformed deposits (1 km away to the E) to the tectonically deformed sandstones close (50 m far) to the OATF. Field data highlighted that the footwall sandstones dips towards W-SW, thus moving towards the OATF, shallower strata progressively crop out, hence from the stratigraphical point of view, porosity should increase due to the decreasing in burial depth. On the other hand laboratory measurements revealed the opposite. Using a permeameter we measured porosity, permeability, and P wave velocity both at ambient pressure and at increasing confining pressure up to 100 MPa, simulating an increase in burial depth up to 4 km. Porosity measured at ambient pressure decreases moving towards the OATF as well as permeability, whilst P wave velocity increased. P wave velocities obtained during depressurization from 100 MPa to ambient pressure were always higher than those recorded during pressurization suggesting inelastic compaction. In order to reconstruct the paleostresses we started from the Athy’s exponential porosity-depth relationship. We calculate the initial porosity at the time of deposition for undeformed sandstones 1 km away from OATF (11.1%) Using stratigraphic and geometrical relationships we calculated that the maximum burial depth of sandstones close to the OATF was about 1500 m. We then calculated that the porosity of sandstones close to the OATF related only to sedimentary load was about 7.4%. This value is higher than the present-day porosity that is 3.7%. The difference ( = 3.7%, equal to inferred porosity minus measured porosity) is thought to be caused by the tectonic load and inelastic compaction associated with the activity of the OATF that changed permanently the petrophysical properties inherited from sedimentation and diagenesis as confirmed by laboratory measurements. The stress needed to reduce porosity from the theoretical value of 7.4% to the measured value of 3.7% at 1500 m depth, is 64.8 MPa. This value represents the maximum differential stress () that acted close to the fault plane (tectonic load). Since field data indicated a compressional regime; this implies that the horizontal stress is 1 and the vertical stress is 3. By using the density-depth relationship, it resulted that, close to the OATF at a depth of 1500 m, 3=37.7 MPa. Consequently, 1, calculated as 1=(3+), is 102.5 MPa. Assuming a coefficient of friction for sandstones of 0.71 and overburden-related inelastic compaction in the proximity of the fault plane, it results that the so calculated stresses are exactly the stress needed to reach critical conditions for slip. Since the OATF has more than 500 m of displacement, critical conditions for slip should have been maintained for long time; this strengthens our methodology that can thus be potentially applied for other tectonically deformed zones.


91


PHYSICAL AND MECHANICAL PROPERTIES OF MAJOR VOLCANIC LITHOLOGIES RELATED TO VOLCANIC STABILITY AND FLANK COLLAPSE:

IMPLICATIONS TO STROMBOLI VOLCANO, ITALY

T Stephens1, S Vinciguerra1,2, R Walker1, P Benson3 N De Paola4 and S Nielsen4

1Geology Department, University of Leicester, Leicester, UK. 2Rock Physics Laboratory, British Geological Survey, Keyworth, UK.

3School of Earth and Environmental Sciences, University of Portsmouth, Portsmouth, UK. 4Department of Earth Sciences, Durham University, Durham, UK.

Volcanic flank collapse is a common process on ocean island and continental volcanoes alike. Volcanic edifices are mainly composed of four lithological groups; lava, autobreccia, pyroclastic material and soil. Although intact lavas are the mechanically strongest components of an edifice, pyroclastic units are more likely to deform at lower stresses and hence localise strain during the early stages of flank collapse. Despite this, physical and mechanical property tests of volcanic material to date focus mainly on lava. To address this, here we report on frictional and mechanical properties of pyroclastic material from Stromboli volcano, Italy.

Stromboli has an unstable NW sector and has experienced four flank collapses in the last 13 ka. Cross-sections of the unstable flank illustrate how variable the group distributions are, both laterally and vertically, with lava accounting for 9–40%, while autobreccia forms 4–20% and pyroclastic material comprises 40–87% of the flank (Apuani et al., 2005; Lucchi., et al 2013). As strain will localise at different stress states within the various lithological groups, it is key to understand the mechanical and physical role of these lithologies with relation to strain localisation.

Competing weakening and strengthening mechanisms will determine the evolution on slip surfaces from low (mm/s) to high slip rates (m/s), leading to large scale flank collapse. Rate and state rotary shear tests have been conducted at low velocity (13 µm/s) to mimic strain localisation within the pyroclastic units under dry and saturated conditions. Further mechanical tests will be conducted using triaxial apparatus and both rock and soil shear box tests to characterise strain localisation and fracture nucleation in intact lithologies as well as the ultimate strength at representative in-situ conditions.


92


FRICTIONAL PROPERTIES OF PHYLLOSILICATE-RICH MYLONITE AND CONDITIONS FOR THE BRITTLE-DUCTILE TRANSITION

Lei Zhang1 and Changrong He1

1State Key Laboratory of Earthquake Dynamics, Institute of Geology, China Earthquake Administration

Phyllosilicate-rich mylonite (60wt. % phyllosilicates which are mainly chlorite and muscovite) is collected from a ductile thrust fault zone along Gengda-Wenmao fault of Longmenshan fault zone. Frictional experiments on artificial mylonite gouge were conducted under elevated temperature in the range of 100–600ºC and effective normal stress of 100 MPa, 200 MPa and 300 MPa to represent the deep portion conditions of the fault zone. In order to obtain velocity dependence of friction, loading rates were stepped up and down in the range of 0.0 4µm/s–1.0 µm/s.

In our experiments, the frictional coefficient of mylonite exhibits systematic increase with increasing temperature. Under 200 MPa and 300 MPa effective normal stress condition, velocity dependence of mylonite gouge shows a transition from initial velocity-strengthening behavior (Regime 1) to velocity-weakening behavior (Regime 2) at about 300ºC and then transitions back to velocity-strengthening behavior (Regime 3) as temperature increases. The velocity dependence of mylonite also shows strong pressure sensitivity. When the effective normal stress is increased to 300 MPa, the stable frictional behavior is significantly enhanced with larger (a-b) compared to that under the lower pressure condition. Microstructure in Regime 3 is characterized by pervasive mylonitic foliations, which is attributed to the plastic deformation of phyllosilicates combined with thermal activated particle size reduction of hard clasts (quartz and plagioclase). At 300 MPa effective normal stress and 400ºC, loading rate exerts an obviously influence on the transition of velocity dependence of frictional strength, from velocity strengthening behavior to velocity weakening behavior as loading rate increases from 0.04 µm/s to 1.0 µm/s.

In the framework of rate and state friction constitutive law, the effects of frictional properties of mylonite on faulting mechanics are discussed. From our experimental results of mylonite, unstable slip events may nucleate in mylonite gouge at 100 MPa effective normal stress and temperatures of 300–600ºC. At higher effective normal stresses of 200 MPa and 300 MPa, the temperature range for the nucleation of unstable slip events narrows, corresponding to 300–500ºC and 300–400ºC, respectively.


93


REACTION ACCOMMODATED CREEP OF WET GABBRO

Yongsheng Zhou1,2, Erik Rybacki2, Changrong He1, Richard Wirth2 and Georg Dresen2

1State Key Laboratory of Earthquake Dynamics, Institute of Geology, China Earthquake Administration, Beijing, China, 1000292GFZ German Research Centre for Geoscience, Potsdam, 14473, Germany

Natural fine-grained gabbro was triaxially deformed in a Paterson-type deformation apparatus in order to evaluate experimentally the flow strength of lower crustal rocks. Composition of the sample was ≈60 vol% plagioclase, ≈30 vol% pyroxene, and ≈10 vol% magnetite and ilmenite. Water content of as-is samples was about 0.15 wt % H2O. We performed creep stepping tests on samples at 300 MPa confining pressure, temperatures between 800°C and 1050°C, and axial stresses of 3–550 MPa, resulting in strain rates between 3x10-4 and 1x10-8 s-1. The mechanical data and microstructural observations indicate that the samples were deformed in the dislocation creep regime. Strain rates decreased with test duration, related to the nucleation and growth of olivine, which is suggested to form by solid-solid reaction in places where opaque minerals are present. The reaction-induced decrease of strain rate can be approximated by an exponential dependence on olivine fraction. Olivine-growth-corrected mechanical data were fitted to a power law creep equation, resulting in a stress exponent of n = 3.0±0.2, an activation energy of Q = 690±25 kJmol-1, and a pre-exponential factor of log A = 15.0±0.3 MPans-1 for wet natural gabbro at T>900°C. Extrapolated to natural strain rates, the fine-grained gabbro is stronger than pure feldspar and weaker than pure pyroxene.


94




11th Euroconference on rock physics and … 11th Euroconference on rock physics and geomechanics 'Holistic rock physics: integrating theory, observation and applications in space and

Documents