TectonoMechanics 2013 - Laboratoire de Géologie de l'Ecole ...

TectonoMechanics 2013Paris | École Normale SupérieureColloquium, April 15-16 th 2013Workshop, April 17-19 th 2013SCIENTIFIC COMMITTEEJérome Fortin, Laboratoire de Géologie, École Normale SupérieureFlorian K. Lehner, Geography and Geology, University of SalzburgYves M. Leroy, Laboratoire de Géologie, Ecole Normale SupérieureBertrand Maillot, Géosciences et Environnement Cergy, Université de Cergy-PontoiseFranz Neubauer, Geography and Geology, University of SalzburgLOCAL ORGANIZING COMMITTEEJérome Fortin, Yves M. Leroy, Bertrand Maillot: tectonomechanics2013@geologie.ens.frMallet Céline mallet@geologie.ens.frChanard Kristel kchanard@caltech.eduAurélien Nicolas nicolas@geologie.ens.frFrancois Passelegue passelegue@geologie.ens.frLucas Pimienta pimienta@geologie.ens.frXiaoping YUAN xyuan@geologie.ens.frORGANIZING PARTNERSFUNDING PARTNERS

ContentsPresentation 2Colloquium Program 3Workshop Program 5Oral Presentations Abstracts 8Poster Presentations Abstracts 311

PresentationTectonoMechanics 2013 keeps with the tradition of the series of colloquium initiated by GeorgMandl years ago in Graz and invites scientists, working in academia or in the industry todiscuss their views on the mechanics of the continental lithosphere and on the physics ofrocks.This edition of TectonoMechanics in Paris will focus on the role of fluids at the scale ofreservoirs and of fold-and-thrust belts and accretionary wedges. The time scales envisionedare related to the earthquake event, the production of reservoirs or the surface transport.The role of heterogeneities in the seismic region of detachements, of compaction and itslocalization in deformation bands – observed in the laboratory or in-situ – in controlling thefluid transfer and pressure are typical questions which could be debated. Discussions of theconsequences of these pressure variations for the production of reservoirs and the spatialand temporal distribution of faulting in frictional wedges are encouraged.The presentations of numerical, physical, and laboratory experiments, or of field workrelated either to regional interpretation or to hydrocarbon exploration and production issuesare thus invited with the only requirement that some aspects of mechanics are debated.Methodological developments towards the quantitative comparison between observationsand theoretical predictions, i.e., the application of inverse problem theory to geological orlaboratory data, are welcome.A two-days workshop will follow, open to engineers, academics and especially graduatestudents, on numerical algorithms for optimization linked to the application of limit analysis.This theory will be presented with illustrations in structural geology and in production relatedproblems, and with demonstrations of numerical codes for frictional and cohesive materials.The benefits of GPU technology, of great interest for optimization in the presence ofnumerous local minima, will be demonstrated. The construction of parallel GPU architecturefor massively parallel computations will be illustrated. This workshop will be animated byparticipants from The University of Newcastle, Australia, The Universidade Nova de Lisboa,Portugal, and the Graduate University of the Chinese Academy of Sciences, Beijing, China,The university of Massachussets at Amherst, USA, the Ecole Normale Supérieure, Paris,and the Université de Cergy-Pontoise, France.2

Colloquium ProgramMonday, April 15 th9h00-9h309h25-9h30CHAIRMAN:9h30-10h0010h00-10h3010h30-11h0011h00-11h3011h30-12h0012h00-12h3012h30-14h30CHAIRMAN:14h30-15h0015h00-15h3015h30-16h0016h00-16h3016h30-17h0017h00-17h3017h30-18h0020h30WELCOME COFFEEIntroduction, Bertrand MaillotJean-Marc DanielThe end of a paradigm? Joints as dilatancy bands rather than mode I fractures,Jean-Pierre PetitFracture propagation in rock with veins: results from DEM models and fieldstudies, Janos UraiTight sedimentary covers for CO 2 sequestration, Dominique LeguillonCOFFEEPermeability-Normal stiffness relationship in shear zones, and scaling laws,Pierre HenryFault reactivation by thermal pressurization of pore fluid: an experimentalevidence, Jean SulemLUNCH & POSTER SESSIONChristophe NussbaumA model of strain localization in porous sandstone as a function of tectonicsetting, burial and material properties; New insights from Provence (SouthernFrance), Roger SolivaYield surfaces and plastic potentials of cemented granular materials from discreteelement simulations, Alfredo TaboadaSimulation of wetting regimes in a granular material, Jean-Yves DelenneCOFFEEFlow structures of granular materials, Émilien AzémaDestabilization process of immersed granular slopes, Patrick MutabarukaParticle finite element method for extreme deformation problems,Kristian KrabbenhoftCONFERENCE DINNER: AU BISTROT DE LA MONTAGNE38 rue de la montagne Sainte-Geneviève - 75005 Paris3

Tuesday, April 16 th8h30-9h00CHAIRMAN:9h00-9h309h30-10h0010h00-10h3010h30-11h0011h00-11h3011h30-12h0012h00-12h3012h30-14h30CHAIRMAN:14h30-15h0015h00-15h3015h30-16h0016h00-16h3016h30-17h0017h00-17h3017h30-18h00WELCOME COFFEEChristian DavidThe effect of cementation in porous carbonate rocks: mechanical data andmicrostructural observations, Jean-Baptiste RegnetMechanical properties and fracturing of carbonate rocks: implication of thesedimentary facies, diagenesis and geodynamics. Example of the Provence(France), Arthur LavenuFluid over-pressure evolution in a foreland basin from diffuse fracture populationat fold-scale, Nicolas BeaudoinCOFFEEThe Coulomb critical taper theory applied to gravitational instabilities,Aurélien LacosteDiscrete modeling of rock avalanches: influence of sizes and shapes of theparticles, Guilhem MollonMass wasting processes on Earth and Mars: gravitational spreading. Numericalsimulations, Magdalena MakowskaLUNCH & POSTER SESSIONMickaele Le RavalecJurassic structural inheritance and Tertiary tectonic reworking along theEurasian margin: insights from the Alpine belt of north Aegean region,Greece-Bulgaria, Nikolay BonevWork minimization can predict timing and path of accretionary Fault Growth,Michèle CookeOn the relationship between forearc deformation, friction properties, porepressure, and megathrust ruptures, Nadaya CubasCOFFEEHydromechanical investigation on a CO2-like altered carbonate: a multiscaleand 3D DIC approach, Louis ZinsmeisterMultiscale laboratory and numerical rock physics - a snapshot of ongoingrocketh-science, Marcel FrehnerCohesive Critical Coulomb Wedges and Rankine Limit States, Florian Lehner4

Workshop ProgramWednesday, April 17 th - Friday, April, 19 thThe workshop is organised as a series of courses and practical exercises. Participants areencouraged to bring their own computer to install softwares that will be provided by thespeakers and to perform the exercises. Preliminary required softwares and preferred OS arelisted below.List of topics and presenters :Basics of Limit AnalysisYves Leroy, ENS, leroyy@geologie.ens.fr. 2 hours.The objective is to present the theory of limit analysis which will be explored from the numericalperspective in other sessions. Several simple examples for extension and compressivefailure using the Coulomb strength criterion will be proposed. The application to the stabilityof fluid satured rocks in accretionary wedges with over-pressures and real cross-sections willbe discussed. Handouts with theory and solution sets will be provided.Computational Limit AnalysisKristian Krabbenhoft, Univ. of Newcastle, Australia Kristian.Krabbenhoft@newcastle.edu.au.3 hours 30.The two main topics are: basic computational limit analysis; and the extension to elastoplasticityincluding dynamics and large deformation. The participants will be able to use abeta-version of a code to test their ability to solve simple problems related to these topics.Computer requirements : the practicals have been tested on Microsoft Windows 7 OS.Sequential Limit AnalysisBaptiste Mary, Univ. de Cergy-Pontoise, France baptiste.mary@u-cergy.fr. 3 hours 30.The sequential limit analysis at the base of the study of the internal deformation of foldand-thrustbelts will be presented. The account of softening, accretion and erosion will bediscussed to show their influence on the timing and the position of fault-thrusting. The fortrancodedevelopped by the author and co-workers for this line of work will be demonstated withsimple examples. Computer requirements : the practicals have been tested on a Linux OS.Necessary packages: gfortran (code compilation), gmt and ffmpeg (post-process visualisation).5

Boundary Integral TechnicsMichèle Cooke, Univ. of Massachussets, USA cooke@geo.umass.edu. 3 hours 30.The objective is to present the fundamentals of the boundary element method for fracturemechanics. The application is the determination of fault propagation via work minimizationfollowing work budget estimation. The beta version of the code developped for this class ofproblems will be presented so that the participants could get their first experience.GPU computationMario J. Vicente Da Silva, Univ. Nova of Lisboa, Portugal, mjvs@fct.unl.pt. 3 hours 30.The use of GPU to speed up optimization problems typical in limit analysis will be presentedwith simple examples ranging from the sequential limit analysis of Baptiste Mary to granularcontact dynamics. The code for a computer with a Tesla card will be presented to illustratethe required programming task.6

Wednesday, April 17 th8h30-9h009h00-11h0011h00-11h3011h30-12h3012h30-14h0014h00-16h0016h00-16h3016h30-18h00WELCOME COFFEEBasics of Limit Analysis, Y. LeroyCOFFEEComputational Limit Analysis 1, K. KrabbenhoftLUNCHComputational Limit Analysis 2, K. KrabbenhoftCOFFEEComputational Limit Analysis 3, K. KrabbenhoftThursday, April 18 th8h30-9h009h00-11h0011h00-11h3011h30-13h0013h00-14h0014h00-16h0016h00-16h3016h30-18h00WELCOME COFFEESequential Limit Analysis 1, B. MaryCOFFEESequential Limit Analysis 2, B. MaryLUNCHBoundary Integral Technics 1, M. CookeCOFFEEBoundary Integral Technics 2, M. CookeFriday, April 19 th8h30-9h009h00-11h0011h00-11h3011h30-13h00WELCOME COFFEEGPU computation 1, V. Da SilvaCOFFEEGPU computation 2, V. Da Silva7

Oral Presentations AbstractsMonday, April 15 th , 9h30-10h00The end of a paradigm? Joints as dilatancy bands rather thanmode I fracturesJean-Pierre Petit 1 and Alexandre Chemenda 21 Géosciences Montpellier, 2 Géoazur, Nice-Sophia Antipolisjean-pierre.petit7@wanadoo.fr ; chem@geoazur.unice.frJoints are the most common fractures in the shallow earth crust. They are essential actorsin brittle deformation and in reservoir properties. Their formation mechanisms have beendebated for over a century (Pollard & Aydin, 1988). The present day widely accepted interpretationis that joints are mode I cracks initiating from defects/flaws and propagating withclearcut separation of the fracture walls due to strong stress and damage concentration atthe fracture tips. In this interpretation, the plumose features (or hackles) are seen as typicalattributes of mode I cracks. They are believed to result from the opening fracture frontbreakdown related to the loading mode change from mode I to mode III (the origin of thischange remains unclear).In GRAM, a synthetic rock analogue material, joint sets (joint wallsbearing plumose marks) were recreated for the first time. It was shown that plumose morphologyis associated with near zero (slightly compressive or tensile) minimal stress duringunloading in experiments starting from relatively high initial mean stress P. Thus the experimentaljoints cannot be mode I fractures. Plumose morphology vanishes with reduction inP, the fracture surface being smooth in purely uniaxial tensile tests when fracture occurs atthe tensile strength. SEM observations of the unopened joints generated in the experimentsunder relatively high P (corresponding to the formation of the plumose marks) reveal thatthese features have a dilation/dilatancy band structure. They represent several grain-thickbands with internal heterogeneous decohesion, with a sinuous band pattern which is relatedto the plumose morphology. This shows that plumose structure is not a signature of openingmode as previously postulated, but of dilatancy bands/joints forming during unloading eventsat near-zero minimal stress. The same SEM scale dilatancy band structures have been evidencedon embryonic joints in jointed dolomicrite, suggesting that natural joints could form asdilatancy bands rather than mode I cracks. Further research into natural dilatancy band/jointstructure coupled with extension experiments in fine-grained rocks is certainly a major challenge.ReferencesChemenda, A. I., S.-H. Nguyen, J.-P. Petit, and J. Ambre (2011), Mode I cracking versus dilatancybanding: Experimental constraints on the mechanisms of extension fracturing, J. Geophys. Res.,116, B04401, doi:10.1029/2010JB008104.Jorand, C., Chemenda, A. I., Petit, J-P., (2012). Formation of parallel joint sets and shear band/fracturenetworks in physical models, Tectonophysics, doi: 0.1016/j.tecto.2011.11.0218

Monday, April 15 th , 10h00-10h30Fracture propagation in rock with veins: results from DEM modelsand field studiesJanos Urai 1 , Simon Virgo 1 , Steffn Abe 1 , Max Arndt 11 Structural Geology, Tectonics and Geomechanics, RWTH Aachen Universityj.urai@ged.rwth-aachen.deOne fundamental aspect of the crack seal process is that an existing vein can act as a heterogeneityin the rock and control the localization of successive fracturing even under constant stress.Observations from crack-seal vein systems suggest that existing veins fundamentally influence fracturepropagation in a rock even in cases where the orientation of the stress field is incompatible withthe orientation of the new fracture.We used a series of 3D Discrete Element Simulations to systematically investigate the influence ofexisting veins with varying orientation and mechanical properties on an approaching fracture. Themodels consist of a tabular heterogeneity within a bonded particle volume fractured under uniaxialtension. The parameters varied in the study are the orientation of the heterogeneity relative to thedirection of uniaxial extension and therefore relative to the orientation of the favourable fracture planeas well as the fracture strength ratio between the matrix material, the vein material and the interfacebetween vein and matrix material.Results show a strong influence of the tabular heterogeneity on the fracture propagation for all orientationsand at cohesion ratios within the range of natural systems. Besides curving and deflection ofthe fracture path associated with changes in fracture mode, bifurcation of fractures as well as arrestof propagation and nucleation of new fractures can be observed. New simulations include repeatedsealing of the fractures and creation of new ones.ReferenceSteffen Abe, Janos Urai, Michael Kettermann Fracture Patterns in Non-Plane Strain Boudinage - Insightsfrom 3D Discrete Element Models. Journal of Geophysical Research. (In press).9

Monday, April 15 th , 10h30-11h00Tight sedimentary covers for CO 2 sequestrationDominique Leguillon 1 , Ekaterina Karnaeva 1,2 , Axelle Baroni 2 , Claude Putot 21 IJLRA CNRS/UPMC, 2 IF,Pdominique.leguillon@upmc.frCO 2 storage at depth is a promising way to reduce the spread of greenhouse gases in the atmosphere.Obviously the sedimentary cover should ensure the sealing of reservoirs. They are naturallyfractured to allow injection and we propose a model to predict the maximum bearable gas pressurebefore reinitiating these fractures in the caprock or incidentally along the interface between thereservoir and the caprock. The method is based on a twofold criterion merging energy and stressconditions (Leguillon, 2002). It has proved its efficiency to simulate rupture tests of v-notched specimensmade of materials like some brittle polymers, ceramics, glass..., as well as to predict the failureof different assembly methods: welding, brazing, gluing. It relies on the joint use of singular elasticfields and matched asymptotic expansions.In the present case, even if the fracture in the reservoir can be compared to a crack, it impinges onthe interface between the reservoir and the caprock and the resulting stress singularity differs significantlyfrom that of a crack in a homogeneous material. It can be stronger or weaker dependingon the elastic contrast between the reservoir and the caprock. In addition to tectonic confinementand overburden load, specific conditions related to the gas pressure acting on the crack faces andthe swelling of the reservoir due to the pressure rise require complementary terms to complete theWilliams expansion that describe the state of stress at the tip of the primary crack.ReferenceLeguillon D. (2002). Strength or toughness? A criterion for crack onset at a notch, Eur. J. of Mechanics- A/Solids, 21, 61-72.10

Monday, April 15 th , 11h30-12h00Permeability-Normal stiffness relationship in shear zones, andscaling lawsPierre Henry 1 , Yves Guglielmi 1 , Frederic Cappa 2 , Pierre Jeanne 11 CEREGE (AMU-CNRS, UMR 7330), Marseille, France2 2GEOAZUR (UNS-CNRS, UMR 7329), Sophia Antipolis, Francehenry@cerege.frDamage to the rock formations around a fault zone as well as cataclasis and chemical alteration inthe fault core generally result in decreased elastic moduli. On the other hand, fault zones may behaveas permeable conduits or impermeable seals depending on the case. Here, we compared thespatial distribution of the permeability and stiffness of two mature fault zones in low-porosity rocksof contrasted geological histories and under different states-of-stress (∼0.1 to 42MPa). Data showan inverse power law relationship between permeability and normal stiffness in the damage zone.A scaling relationship between fracture aperture and fracture length with an exponent of about 0.85can explain the relationship found. This behavior may be understood if the fault zone is consideredas a self-affine object, with an exponent in the same range as reported for the roughness exponentof individual fractures. However, the scaling relationship between aperture and shear displacementon individual fractures is found to have a low exponent (about 0.3) compared to that expected forrigid sliding of self-affine fracture surfaces. This suggests other processes such as asperity crushing/abrasionand cementation contribute to the evolution of shear fracture properties with slip. Inaddition, fault mirrors are anisotropic, their roughness exponent being lower in the direction parallelto fault slip than in the perpendicular direction (e.g. Renard et al., 2006). It follows that the exponentdetermining the opening of a fracture as a function of slip could be different from the exponent controllingthe scaling between hydraulic aperture and stiffness. Eventually, a continuous layer of gougematerial forms, and this core layer determines the normal stiffness of mature fault zones. Dependingon the amount of clay material, the permeability may be decreased by several orders of magnitude ormay fit the scaling relationship. This variability in poroelastic properties and permeability implies thatdifferent coupling between fluid pressure and strain exists in the fault core and damage zone.11

Monday, April 15 th , 12h00-12h30Fault reactivation by thermal pressurization of pore fluid: an experimentalevidenceJ. Sulem 1 , M. Monfared 1 , P. Delage 11 Laboratoire Navier-CERMES, Ecole Nationale des Ponts et Chaussées, IFSTTAR, CNRS, UniversitéParis-Est, Marne-la-Vallée, FranceJean.Sulem@enpc.frUndrained heating of saturated geomaterials leads to pore fluid pressurization, because the thermalexpansion coefficient of water is much higher than the one of the solid matrix. This is enhanced inthe case of clayey soils and rocks because of the possibility of thermal collapse of clay. Vardoulakis,(2002) and Sulem et al (2007) have demonstrated the importance of thermally collapsible and thermallysoftening clay on the overall dynamic thermo-poro-mechanical behaviour of clay-rich layers forlandslides and of clay-rich gouges for seismic slip. Pore fluid pressurization inside the slip zones leadto substantial reduction of the apparent friction which leads to an increase of the slip rate. This phenomenonis reproduced experimentally on a Boom clay sample which is previously sheared in orderto form a shear band and then heated in undrained conditions. It is obtained that the failure occursalong the existing shear band which behaves as a weakness plane in the sample. The responses ofthe local strain measurement transducers clearly show the sliding of rigid blocks when failure occurs.The effect of the presence of a shear band on the permeability of the other Boom clay sample isinvestigated at ambient temperature and at 80 o C (Monfared et al, 2012). It is shown that the presenceof a shear band does not affect significantly the permeability. These results confirm the goodself sealing properties of Boom clay at ambient and at high temperature. It also corroborates otherpublished results on the favourable properties of the clay host rock for the containment and isolationof nuclear waste in that sense that shear fractures in the damage zone around excavations do notaffect significantly the hydraulic conductivity (Armand et al., 2013).ReferencesArmand, G., Leveau, F., Nussbaum, C., Vaissiere, R., Noiret, a., Jaeggi, D., Landrein, P., et al. (2013).Geometry and Properties of the Excavation-Induced Fractures at the Meuse/Haute-Marne URL Drifts.Rock Mechanics and Rock Engineering.Monfared, M., Sulem, J., Delage, P., & Mohajerani, M. (2012). On the THM behaviour of a shearedBoom clay sample: Application to the behaviour and sealing properties of the EDZ. Engineering Geology,124, 47-58.Sulem, J., Lazar, P., & Vardoulakis, I. (2007). Thermo-poro-mechanical properties of clayey gougeand application to rapid fault shearing. International Journal for Numerical and Analytical Methods inGeomechanics, 31(3), 523-540.Vardoulakis, I. (2002). Dynamic thermo-poro-mechanical analysis of catastrophic landslides. Géotechnique,52(3), 157-171.12

Monday, April 15 th , 14h30-15h00A model of strain localization in porous sandstone as a functionof tectonic setting, burial and material properties; New insightsfrom Provence (Southern France)Roger Soliva 11 University of Montpellier II, Laboratoire Géosciences Montpellier, Montpellier, Francesoliva@gm.univ-montp2.frThe analysis of three cataclastic band sets from Provence (France) reveals that the band density,their conjugate angles, their ratio of shear displacement to compaction, and the amount of cataclasiswithin the bands differ and can be expressed as functions of tectonic setting and petrophysical properties.We identify (1) a dense and closely spaced network of shear-enhanced (reverse) compactionbands; (2) a regularly spaced less dense network of reverse compactional shear bands; and (3) alocalized network of normal shear bands. The field data show that strain localization is favored in anextensional regime and is characterized by shear bands with a large shear to compaction ratio and asmall conjugate band angle. In contrast, distributed strain is favored in a contractional regime and ischaracterized by compactional bands with a low ratio of shear to compaction and a large conjugateband angle. To explain the mechanical origin of this strain localization, we quantify the yield strengthand the stress evolution in extensional and contractional regimes in a frictional porous granular material.We propose a model of strain localization in porous sands as a function of tectonic stresses,burial depth, material properties, strain hardening and fluid pressure. Our model suggests that stressreduction, inherent to extensional regime, favors strain localization as shear bands, whereas stressincrease during contraction favors development of compactional bands.ReferenceRoger Soliva, Richard A. Schultz, Gregory Ballas, Alfredo Taboada, Christopher Wibberley, ElodieSaillet, Antonio Benedicto , 2013. A model of strain localization in porous sandstone as a function oftectonic setting, burial and material properties; new insight from Provence (southern France), Journalof Structural Geology, 49, 50-63.13

Monday, April 15 th , 15h00-15h30Yield surfaces and plastic potentials of cemented granular materialsfrom discrete element simulationsNicolas Estrada 1 and Alfredo Taboadab 21 Departamento de Ingenieria Civil y Ambiental, Universidad de los Andes, Bogota, Colombia2 Laboratoire Géosciences Montpellier, UMR 5243, Université de Montpellier 2-CNRS, Place EugèneBataillon, F-34095 Montpellier cedex 5, Francen.estrada22@uniandes.edu.co; alfredo.taboada@gm.univ-montp2.frWe present a numerical investigation on the mechanical behavior of cemented granular materials bymeans of discrete element simulations. In particular, two conceptual elements of this behavior wereexplored: the yield surfaces and the plastic potentials, as defined in the framework of elastoplasticconstitutive models. For this purpose, we implemented a contact model that mimics the mechanicalbehavior of cemented bonds, we constructed a set of polydisperse samples with different solidfractions, and we tested these samples by means of a biaxial device with four rigid walls. The yieldsurfaces and plastic potentials were thus obtained and analyzed as functions of the solid fraction ofthe material.We found that the yield surfaces are smooth closed curves with two axes of symmetry: the meanstress axis and an axis that is parallel to the deviatoric stress axis. For loose materials, the yieldsurfaces are well fitted using two functional forms: an ellipse and a rugby ball shaped curve. As thedensity of materials increases, the yield surfaces becomes bigger, their aspect ratio increases, andtheir center shifts to larger values of mean stress. In the limit case of the densest material studied,the yield surface becomes an open curve, whose asymptotes can be understood as the Coulombstrength envelope for a cohesive material.The plastic potentials can be represented by a family of curves whose shape is similar to that of theyield surfaces. For the loosest materials studied, the flow rule is close to being associated. For densermaterials, the flow rule gradually deviates from normality. These results confirm previous experimentalfindings, evidencing the great potential of discrete element simulations in exploring the behavior ofreal cemented geomaterials.ReferenceEstrada, N. and A. Taboada, 2013, Study of the Yield Surface and the Plastic Potential of CementedGranular Materials using Discrete Element Simulations, Computers and Geotechnics, 49, 8 pp.14

Monday, April 15 th , 15h30-16h00Simulation of wetting regimes in a granular materialJean-Yves Delenne 1 , Vincent Richefeu 2 , Farhang Radjai 31 NRA-IATE, Montpellier2 2MGC CNRS-Univ. Montpellier 23 S-R Univ. J. Fourier, Grenobledelenne@supagro.inra.frGranular materials such as soils are often filled by a wetting or binding liquid that plays a key role inthe rheological properties. For example, the triggering of landslides involves the rheology of densemixtures of grains and water; the presence of methane-hydrates at specific conditions of depth andtemperature endows the marine sediment a mechanical strength... Another example is the so-calledvadoze zone where water is retained by adhesion and capillary action without being necessarily saturated.The properties of this zone, which is above the phreatic zone, are essential for agriculture andpollution transport.In this talk we present a numerical study of the distributionof liquid clusters and the evolution of capillary cohesion in2D granular assemblies with increasing liquid volume condensedfrom a homogeneously injected vapour phase. Thethermodynamics of phase change is based on Carnahan-Starling’s equation of state from which the interactions betweenliquid, gas and solid (grains) are derived using nonlocalpotentials in the framework of Multiphase Lattice Boltzmannmethod. A percolation algorithm is employed to identifythe liquid clusters and to determine their volume and connectivitywith the surrounding grains. The internal pressuresof the clusters is analysed as a function of liquid content. Weare thus able to obtain the global liquid-retention curve aswell as the forces acting on the grains. Finally, we computethe negative pressure in the sample due to capillary forces.We show that the evolution of this negative pressure with theamount of liquid reveals four different states reflecting the connectivity of the liquid phase and localgrain environments. Our simulations reproduce correctly the classical relation between the Laplacepressure and the amount of liquid. We find that the liquid phase undergoes a percolation transitionfor a liquid content well below full saturation. Interestingly, the cohesive strength has its peak valuebelow this transition, dividing thus the funicular regime into an ascending cohesion regime followedby a descending cohesion in the late funicular regime.ReferencesX. Shan and H. Chen, Lattice Boltzmann model for simulating flows with multiple phases and components,Physical Review E, 47(3): 1815-1820, 1993. M. Scheel, R. Seemann, M. Brinkmann, M. DiMichiel, A. Sheppard, B. Breidenbach, and S. Herminghaus. Morphological clues to wet granular pilestability. Nature Materials, 7(3):189-193, 2008.15

Monday, April 15 th , 16h30-17h00Flow structures of granular materialsEmilien Azéma 1 and Farhang Radjai 11 LMGC, Université Montpellier 2, CNRSemilien.azema@univ-montp2.fr and franck.radjai@univ-montp2.frBy means of three-dimensional Contact Dynamics (CD) simulations, we investigate the microstructuralproperties of sheared granular packings by varying systematically the inertial number I from thequasi-static regime, to the intermediate dense flow and collisional regime (Gdr Midi, 2004). A detailedanalysis of the connectivity of the particles reveals that the transition between the regimes iscorrelated to profound changes of the nature of the contact network. In particular, we show that theincrease of friction angle with inertial number is only a consequence of contact anisotropy. As theinertial number increases, the role of force chains decrease and that of friction mobilization increasesas long as percolating force chains are present in the system. Beyond this point, which correspondsto an inertial number I 0 ∼ 0.25, the force anisotropies do not evolve with the inertial number. Finally,a detailed analysis of the spatial correlations of floating particles, i.e., particles without bearing contacts,reveals the occurrence of "fluidized’" volumes in the dense flow regime which increase in sizeand percolate precisely at I 0 as illustrated in the following figures.ReferenceGdr Midi, (2004), On dense granular flows, Eur. Phys. J. E, 341-365.16

Monday, April 15 th , 17h00-17h30Destabilization process of immersed granular slopesPatrick Mutabaruka 1 , Jean-Yves Delenne 2 , Kenichi Soga 3 , Farhang Radjai 41 , 4LMGC CNRS-Univ. Montpellier 22 INRA-IATE, Montpellier3 Cambridge Universitypatrick.mutabaruka@univ-montp2.frWe analyze the destabilization process of an inclined granular bed immersed in a viscous fluid bymeans of 3D coupled molecular dynamics/Lattice Boltzmann simulations. Extensive simulations wereperformed for different values of the packing fraction, slope angle and liquid viscosity. Above a criticalpacking fraction, we observe a slow creep deformation as a result of negative pore pressures inducedby dilatancy, as also previously evidenced by experiments (Pailha et al., 2008). We find that the slopefailure is initiated while the packing fraction is above the critical packing fraction (0.58 for sphericalparticles of nearly the same size), and the total creep deformation before failure increases with packingfraction. The dilation of the bed and liquid flow in the pore space are strongly inhomogeneous.But the pore pressure gradients are consistent on average with a simple model based on the quasiequilibriumstate of the bed and Darcien drag within the bed. We show that the observed featuresof the destabilization process may be predicted from the conjecture that, in contrast to the hardeningplastic behaviour of granular materials, the strength of the bed under a constant stress ratio due tothe gravity declines as the current packing fraction decreases. Hence, the slope is destabilized for apacking fraction at which the shear strength falls below the shearReferenceM. Pailha, M. Nicolas and O. Pouliquen, Physics of Fluids 20, 111701 (2008).17

Monday, April 15 th , 17h30-18h00Particle finite element method for extreme deformation problemsKristian Krabbenhoft 11 Centre for Geotechnical Science and Engineering, University of Newcastle, NSW, Australiakristian.krabbenhoft@newcastle.edu.auThe numerical solution of very large deformation problems involves two main challenges: i) meshdistortion as a result of large changes in geometry and ii) solution of the discrete governing equationsin an efficient and robust manner.The first challenge, which is implicitly connected to the conventional finite element method, has motivatedthe formulation of a large number of meshless methods that in one way or another alleviate theproblems faced with mesh distortion in the finite element method. However, these methods introducecomplications of their own, both practical and fundamental, and are from a mathematical point ofview much less established than the finite element method. The second challenge, that of solving thediscrete governing equations in an efficient and robust manner, is to a large extent independent ofmethod of discretization. In the context of the finite element method this is an issue even for infinitesimaldeformation problems and extending the scope of deformation beyond the infinitesimal rangeobviously does not improve the situation. These considerations have motivated the development of anew scheme that addresses both issues explicitly. The first point is addressed via the so-called ParticleFinite Element Method (PFEM) originally proposed by Onate and his co-workers for the solutionof fluid dynamics problems involving free surfaces. The idea is the following: in each time step a standardLagrangian formulation of the governing equations is discretized using the conventional finiteelement method. After solving, the coordinates of the nodes are updated according to the displacementsâĂŞ without any consideration for the mesh. One is then left with a âĂIJcloud of particlesâĂİ(the nodes) on the basis of which the corresponding solid is identified. This is done using the alphashapemethod and once identified, the solid is again discretized using the conventional finite elementmethod to set up the discrete governing equations for the next time step.The second point is addressed by state-of-the-art mathematical programming methods that guaranteea robustness not offered by conventional Newton-Raphson based procedures for nonlinear finiteelement analysis. The end result is a method capable of handling very large deformations with a reliabilitythat makes it suitable for general purpose use. In my talk I will present example of applicationto various engineering problems involving off-shore oil and gas infrastructure as well as to variousearth science applications.18

Tuesday, April 16 th , 9h00-9h30The effect of cementation in porous carbonate rocks: mechanicaldata and microstructural observationsRegnet J.B. 1 , David C. 1 , Robion P. a, Fortin J. 2 , Collin P.Y. 31 Univ. de Cergy-Pontoise et Env. Cergy, 5 Mail Gay-Lussac, F-95031 Cergy-Pontoise cedex, France2 Laboratoire de Géologie, Ecole normale supérieure, CNRS UMR 8538, Paris, France3 UPMC Univ. Paris 06, UMR 7193, ISTeP, Paris, Francejean-baptiste.regnet@u-cergy.frThe contribution of early cements to the lithification of thin rock layers is a well-known phenomenon inlimestone. Early cementation limits mechanical compaction, creating a heterogeneous medium fromthe earliest stages of diagenesis (non-touching grains, preservation of intergranular porosity...). Suchcements are common in shallow carbonate platforms developed during calcite sea periods (Sandberg,1983) and have been well-illustrated from the Middle Jurassic of the Paris Basin by Purser(1969). In recent studies, P-wave velocity measurements performed on samples from the OolitheBlanche formation (Middle Jurassic Limestone, Paris basin) show that dynamic moduli were largelycontrolled by microstructural arrangement of both porosity and cement around or within the ooids(Casteleyn et al., 2010, 2011). In order to investigate the effect of such cement on the mechanicalbehaviour of limestone, we deform in conventional triaxial configuration two carbonate samplespreviously investigated by Casteleyn et al., 2010, 2011 (called respectively Ra02 and By04). Bothdisplay comparable porosity and permeability values (respectively 18% and 0.2 mD). Samples havesimilar facies and composition (oolithic microporous grainstone with a 500Âţm average grain size).Petrographic observations indicate that Ma02 is affected by an early diagenetic stage with isopachousfibrous cement around grains whereas By04 does not display such fringes. Triaxial experiments wereperformed on both samples under saturated conditions with a 0.5 MPa pore pressure and a 1.5 MPaconfining pressure. Two different mechanical behaviours were observed:- For the early cemented sample (Ra02): a typical behaviour of the brittle failure regime, with stressstraincurves reaching a peak (20 MPa) beyond which strain softening was recorded, and strainlocalization on a shear fracture.- For the early compacted sample (By04): a more ductile behaviour beyond the yield point (24 MPa)and no localization on a shear fracture.Petrographic observations after deformation reveal a large proportion of damaged grains in the brittlesample (Ra02), with spalling of oolite borders and isopachous fringes. Damage was distributed homogeneouslyin the ductile sample (By04). Our set of data suggests that the amount of isopachousfibrous cement seems to be a microstructural parameter which has a significant control on the mechanicalbehaviour in carbonate rocks.ReferencesCasteleyn L., Robion P., Collin P.Y., Menendez B., David C., Desaubliaux G., Fernandes N., DreuxR., Badiner G., Brosse E., Rigollet C., 2010 - Interrelations of the petrophysical, sedimentological andmicrostructural properties of the Oolithe Blanche Formation (Bathonian, saline aquifer of the ParisBasin). Sedimentary Geology 230 (2010) 123-138.Casteleyn L., Robion P., David C., Collin P.Y., Menendez B., Fernandes N., Desaubliaux G., RigolletC., 2011. An integrated study of the petrophysical properties of carbonate rocks from the “OolitheBlanche” formation in the Paris Basin. Tectonophysics 503 (2011) 18-33.Purser B.H., 1969 - Syn-sedimentary marine lithification of Middle Jurassic in the Paris Basin. Sedimentology12, 205-230.Sandberg P.A., 1983 - An oscillating trend in Phanerozoic non-skeletal carbonate mineralogy. Nature305, 19-22.19

Tuesday, April 16 th , 9h30-10h00Mechanical properties and fracturing of carbonate rocks: implicationof the sedimentary facies, diagenesis and geodynamics.Example of the Provençale (France) and Apulian (Italy) platformsArthur P.C. Lavenu 11 Aix-Marseille Université / Totalarthur.lavenu@total.comIn some way, all hydrocarbon reservoirs are affected by natural fractures. Hence, characterizingfracture networks in Naturally Fractured Reservoirs (NFRs) is a major challenge for hydrocarbon explorationand production. The fracture organization needs to be deciphered as it is the key parameterfor the understanding of fluid flow, as fractures control most of the permeability through the reservoir.However, at present-day, fracture prediction in the subsurface is poorly constrained becauseof lack of understanding of controls on fracture occurrence. Fracture parameters needed for betterfluid flow comprehension and modeling (such as aperture, connectivity, extension, etc.) are only providedthrough seismic, well imaging and cores, but these data remain not sufficient to define the 3Ddistribution of a fracture network. Moreover, fracture characterization within a carbonate reservoir isstrongly dependent on the mechanical properties of the host-rock, and thus, is only possible throughthe close study of the relationship with host-rock heterogeneities. Therefore, studying fractures inoutcrop reservoir analogues is a way to access the spatial complexity of fracture patterns and todecipher the mechanical properties of rocks and their impact on fracture distribution. There, the conditions(origin, nature, geometric parameters...) for the fracturing process through the geodynamichistory of the reservoir can be determined. To this purpose, we have targeted platform carbonatesfrom different places and of different age that have undergone different geodynamics. We aim todetermine the sedimentary and diagenetic controls on fracture patterns and the genetic correlationwith tectonic and burial history. We analyzed fracture patterns and rock facies as well as the tectonic,diagenetic and burial history for each site. Fracture patterns are determined from geometrical, kinematicand diagenetic criteria based on field and lab measurements. Fracture sequences are definedbased on cross-cutting and abutting relationships and compared with geodynamic history and subsidencecurves. In both areas we observe a stage of fracturing perpendicular to bedding, synchronouswith shallow burial and prior to major tectonic events. Fractures are joints and veins and numerousbed-parallel stylolites which densities can be correlated to rock porosity. Carbonates have undergoneearly diagenesis during fast and shallow burial, conferring early brittle behavior. Subsequently, themechanical bedding was acquired during early burial, controling fracture pattern and its development.The amount of stylolites is not correlated to burial depth but to fracture density, porosity and free airP-wave velocity. It appears that fractures development, mechanical and petrophysical properties areacquired during early diagenesis.20

Tuesday, April 16 th , 10h00-10h30Fluid over-pressure evolution in a foreland basin from diffusefracture population at fold-scaleNicolas Beaudoin 11 ISTeP, Université Paris 6nicolas.beaudoin@upmc.frDeciphering the evolution of fluid pressure during fold history is a challenging problem because onlyfew methods exist to constrain its evolution through time, especially when no hydrocarbons are encounteredin the paleo(fluid)system. This contribution aims at presenting and discussing a new approachto reconstruct the evolution of fluid (over-)pressures based on paleostress analyses (Amrouchet al., 2011). Our study focuses on limestones of the famous Mississippian-Permian Madison andPhosphoria formations in Laramide basement-cored folds of the Bighorn Basin (Wyoming, USA).The combination of stress inversion of fault slip and calcite twin data with rock mechanics allows determiningboth orientations and magnitudes of principal stresses during basin evolution. Assuming noburial change through the Layer Parallel Shortening (LPS) phase, the comparison of the determinedeffective vertical stress with its reference value calculated with respect to the paleo-overburden andhydrostatic fluid pressure) may be used to quantitatively estimate fluid overpressure at different stepsof the tectonic history. Alternatively, if hydrostatic fluid pressure is assumed to prevail, anomalousvertical stresses likely reflect overburden variations. The application focuses on the diffuse fracturepopulations observed at both the Sheep Mountain and the Rattlesnake Mountain anticlines to constrainfluid overpressure variations during both the Sevier (thin-skinned) and Laramide (thick-skinned)tectonic phases, at both fold and basin scale.Results highlight an initial fluid overpressure in limestones buried at 2 to 3 km depth and emphasizethat the LPS-related stress build-up during Sevier and Laramide is associated with an increase influid overpressure until the lithostatic value is reached. In contrast, we identified peculiar periodsduring which fracture sets developed under a nearly hydrostatic fluid pressure, suggesting a highhydraulic permeability of the sedimentary cover due to their tensile nature, a conclusion supportedby independent geochemical studies on the vein cements and by field observations. At the basinscale, the evolution of the fluid overpressure likely reflects the eastward fluid migration in the RockyMountain foreland during the Sevier and Laramide contractions. Finally, considering fluid overpressurereleased during folding permits to estimate syn-folding exhumation of strata, the value of whichis consistent with independent paleo-barometric reconstructions based on hydrocarbons fluids andon exhumation rate reconstructed from fission track studies.ReferenceAmrouch, K., Beaudoin, N., Lacombe, O., Bellahsen, N., Daniel, J.-M. (1992), Paleostress magnitudesin folded sedimentary rocks, Geophysical Research Letters 38, L17301.21

Tuesday, April 16 th , 11h00-11h30The Coulomb critical taper theory applied to gravitational instabilitiesAurélien Lacoste 1 , Régis Mourgues 2 and Cynthia Garibaldi 21 E.A.6293 GéHCO, Université de Tours, Parc Grandmont, 37200 Tours2 LUNAM, LPGN - UMR6112, Université du Maine, 72085 Cedex 9aurelien.lacoste@univ-tours.frThe role of pore-fluid pressure on the evolution of deformation has been demonstrated by numerousmechanical tests and theoretical studies, and largely investigated in the scope of the understandingof the mechanics of fold and thrust belts and accretionary prisms. In particular, Davis et al. (1983),and a series of subsequent studies, showed that the mechanical and structural evolution of wedgeshapedmasses deforming above a detachment plane could be described by the Critical CoulombWedge theory.The exact solutions for noncohesive wedges provided by Dahlen (1984) express the critical taper (i.e.,the surface slope angle + the basal dip) as a function of angles between the maximal principal stress,the surface of the wedge and the basal detachment. Where the basal shear stress is directed towardsthe toe of the wedge, the first solution corresponds to a compressive wedge. This solution has beenthe most frequently studied case and was widely supported by natural, experimental and numericaldata. However, few studies deal with the applicability of the other solution to gravitational spreadingand gliding of downslope-thinning wedges along passive margins, where elevated pore-fluid pressureis common in sediments.On the basis of the work by Dahlen (1984), we derived an alternate formulation better suited to thestudy of gravitational spreading of a system subjected to pore-fluid pressure. Our solution allows fora direct calculation of the fluid pressure (or the basal friction) required for sliding on the base. Inour derivation, we corrected the influence of pore pressure on the basal detachment, and modifiedthe expression of the basal effective friction. On one hand, this correction has no influence on thesolution applicable to compressive wedges. On the other hand, it slightly increases the basal porepressure required for the triggering of gravitational instabilities.To verify the predictions of the theory, we carried out a series of physical experiments in which compressedair was used to trigger sand wedge instabilities. The models were built on a sieve andcomprised a basal low-permeability glass microbeads layer, overlain by a coarse sand cover. Thehigh permeability contrast between the glass microbeads and the sand allowed for fluid overpressureat the base of the models. The coefficient of fluid pressure was constant along the base of the models.We measured, for different values of the taper of the wedge, the fluid pressure required to triggersliding. The experimental results are in good agreement with our analytical predictions.ReferencesDavis, D., Suppe, J., Dahlen, F.A. (1983), Mechanics of fold-and-thrust belts and accretionary wedges,Journal of Geophysical Research 88(B2), 1153-1172.Dahlen, F.A. (1984), Non-cohesive critical Coulomb wedges: an exact solution. Journal of GeophysicalResearch 89, 10125-10133.22

Tuesday, April 16 th , 11h30-12h00Discrete modeling of rock avalanches: influence of sizes andshapes of the particlesMollon Guilhem 1 , Richefeu Vincent 1 , Daudon Dominique 1 , Villard Pascal 11 3SRLab, CNRS UMR 5521, UJF Grenoble 1, Grenoble-INPguilhem.Mollon@insa-lyon.frRock avalanches are a significant concern in developing mountain areas. Thus a reliable predictionof depositional areas from avalanches is needed. In order to improve the numerical modelling ofsuch events and to provide information concerning the physical phenomena underlying this type ofgranular flow, a discrete element model, which takes into account frictional and collisional dissipationat grain scale together with angular-shaped elements, is used to investigate the collective behaviorof granular masses propagating down a slope. The discrete element model (DEM) parameters aredefined from drop tests involving the collision of an individual particle with a flat surface (Richefeuet al., 2012). The validity of the numerical model is estimated by comparison with the results of alaboratory experiment from the literature, involving a dry granular flow on an inclined plane. Thenumerical model improves the understanding of rock avalanches by providing both valuable informationabout the way energy is dissipated either at the base or within the propagating granular massand relevant information about the kinematics of the flow and the shape of the deposit (Mollon et al.,2012). The influence of contact-law parameters is investigated using a sensitivity study. It is shownthat the flow is strongly influenced by basal friction, while inter-particle friction and collisional dissipationphenomena intervene mostly in areas of flow perturbation (such as transition zones betweentwo slopes). A macroscopic roughness of the slope surface induces an increased disorder in theparticle motion which increases both frictional and collisional dissipation within the granular mass.Using a planar slope and increasing the frictional parameter can only partly reproduce the apparentinfluence of this roughness. The last section is dedicated to the close study of the influence ofcritical parameters, such as the sizes and shapes of the particles or the total volume of the avalanche.ReferencesRichefeu, V., Mollon, G., Daudon, D., and Villard, P. (2012), "Dissipative contacts and realistic blockshapes for modelling rock avalanches", Engineering Geology, 19-150 (2012), 78-92,Mollon, G., Richefeu, V., Daudon, D, and Villard, P. (2012), "Numerical simulation of rock avalanches:Influence of local dissipative contact model on the collective behaviour of granular flows", Journal ofGeophysical Research, Solid Earth, AGU, 117(2012)23

Tuesday, April 16 th , 12h30-12h30Mass wasting processes on Earth and Mars: gravitational spreading.Numerical simulationsMagdalena Makowska 1 , Frederic Gueydan 2 , Daniel Mege 11 Institute of Geological Sciences, Polish Academy of Sciences, Research Centre in Wroclaw, Poland2 Géosciences Montpellier, UMR5243,Université de Montepellier, Francemagda.m.makowska@gmail.comDeep-seated gravitational slope deformation (DS-GSD) has been interpreted in Valles Marineriscanyon on Mars and also documented on Earth.Diagnostic feature for this processes are: dublecrest, uphill - facing scarps, trench, counterscarpsand downslope bulging morphology (Fig1). Obtained scarp dimension on Mars havebeen compared with scarp dimensions at textbookwith terrestrial analogues in the PolishTatra Mountains. The vertical offset of faultscarps does not exceed 10 meters, whereasin Valles Marineris fault offset is 40-1000 meters.Figure 1: Fig.1 DSGSD on Earth and analogue featureson Mars. a,b views of crestal graben and uphill- facing scarps in Geryon Montes, Valles Marineris,Mars (CTX image). c, d views of this same structureson Earth in the Tatra Mountains, Poland nearKondracka Kopa and Wolowiec in the Slovakian partof Tatra Mountains. (photo c,d by K. Debniak).In both instances, DSGSD has been interpreted tobe a postglacial process. DSGS triggering mechanismand strain localization in the terrestrial andMartian cases are investigated and compared usingthe finite element code SARPP. The constraints introduced into the model are presented in fig.2. The internal part of the ridge is modelled using "weakening points" that reflect the influence of theglacier erosion on the slope stabilityReferencesBachmann, D., Bouissou, S., Chemenda, A.(2009), Analysis of massif fracturing during Deep-Seated Gravitational Slope Deformation by physicaland numerical modelling, Geomorphology,103, 130-135.Kromuszczyńska, O., Mege, D., Lucas, A., Gurgurewicz,J. (2010), Giant sacking scarps in VallesMarineris, abstract.Figure 2: Principle of finite element modeling of DS-GSD. Initial state of stress in the topographic ridge (a)is calculated (b) as a function of ρ-density, g-gravity, ζ-state of stress ρ ice density ρ rock density.24

Tuesday, April 16 th , 14h30-15h00Jurassic structural inheritance and Tertiary tectonic reworkingalong the Eurasian margin: insights from the Alpine belt of northAegean region, Greece-BulgariaNikolay Bonev 11 Sofia University St. Kliment Ohridskiniki@gea.uni-sofia.bgJurassic ophiolites in the eastern Vardar suture zone in Chalkidiki peninsula of northern Greeceformed in NW-SE oriented wrench corridor (Bébien et al., 1986) recording rift-spreading evidencedby NE-SW trending sheeted dykes. Obduction of the Vardar zone ophiolite onto the Circum-RhodopeBelt (CRB) and Serbo-Macedonian-Rhodope terrane (Eurasian margin) involved Late Jurassic NEdirectedthrust and dextral strike-slip displacement localized in ENE-SSW oriented shear zones andfollowed by Tertiary dextral strike-slip overprint. Counterpart CRB Jurassic Evros ophiolite in Thraceregion of north-eastern Greece shows sheeted-like dykes trending NW-SE to NE-SW and northwardthrust direction, and is now bounded by Eocene-Present Thrace basin. NE-SW oriented extension inthe north Aegean region (Bonev et al., 2010 and references therein) overlapped with the Late Eoceneto Miocene 41 ◦ clockwise rotation of the Chalkidiki Peninsula (e.g. Brun and Sokoutis, 2007) and thedevelopment of fault-controlled South Rhodope Basin System (Koukouvelas and Aydin, 2002) linkedto the North Anatolian Fault System in Thrace region, where no rotation occurred. When restoredto the pre-Eocene rotation, both Chalkidiki peninsula and Thrace region are aligned across the northAegean, with similar tectonic patterns of Jurassic and Tertiary faults and shear zones. Therefore,based on structural geometry and kinematics, the Tertiary lower crust ductile flow and brittle uppercrust stretching and strike-slip movement had reworked tectonic grain inherited from the Jurassicevent that has created mechanical heterogeneity in the crust.Acknowledgments: This work has been supported by NFSR, grant no. DDVU 02/94.ReferencesBébien, J., Dubois, R., Gauthier, A. (1986), Example of ensialic ophiolites emplaced in a wrenchzone: Innermost Hellenic ophiolite belt (Greek Macedonia), Geology, 14, 1116-1119.Brun, J-P., Sokoutis, D. (2007), Kinematics of the Southern Rhodope Core Complex (North Greece),International Journal of Earth Sciences, 96, 1079-1099.Bonev, N., Spikings, R., Moritz, R., Marchev, P. (2010), The effect of early Alpine thrusting in latestageextensional tectonics: Evidence from the Kulidzhik nappe and the Pelevun extensional allochthonin the Rhodope Massif, Bulgaria, Tectonophyscics, 488, 256-281.Koukouvelas, K.I, Aydin, A. (2002), Fault structure and related basins of the north Aegean Sea andits surroundings, Tectonics, 21(5), 104625

Tuesday, April 16 th , 15h00-15h30Work minimization can predict timing and path of accretionaryfault growthMichèle L. Cooke 1 , Justin W. Herbert 1 , Jessica Mc Beck 1 and Bertrand Maillot2 21 Univ. of Massachusetts2 Univ. of Cergy-Pontoisecooke@geo.umass.eduFault systems may evolve to minimize the total work on the system (e.g. Cooke et al, in press). Thiswork budget of fault systems includes tectonic work, frictional heating, internal work, work againstgravity and seismic energy (Cooke & Murphy, 2004; Del Castello & Cooke; 2007). Often overlookedis the work required to propagate faults. While this term may be smaller than the other terms in thework budget, observations of long-lived faults suggests that Wprop is not negligible. Furthermore, thisterm controls the timing of new fault growth. Measurements of external work from sandbox experimentsat UCP show a drop in work associated with the growth of each fault (e.g. Souloumiac, 2012).The new faults develop when the total work savings due to slip along the fault exceeds the cost ofgrowing the new fault (Del Castello and Cooke, 2007). We use the drop in work associated with newfault formation in the UCP experiments to estimate that new faults cost 0.15 J/m2. This suggests thatduring the episodes of no new faulting, the potential work savings of growing a new fault does notexceed 0.15 J/m2. We investigate episodes of fault growth within 2D Boundary Element Method numericalsimulations of the UCP experiments. Stress and PIV data collected from the experiments areused to determine material properties. Three models simulate the conditions during 1) initial loadingof the sand, 2) just prior to new fault growth and 3) after the forethrust propagates to the sandbox surface.The changes in external work between these models parallel the measured changes in force.To explore the potential fault growth during loading and just prior to the observation of new faulting weimplement a new code that searches for the fault growth path that minimizes external work. For themodel that simulates initial loading of the sandbox, the new code may predict fault growth but the costof producing this fault exceeds the benefit of having the fault. In contrast, the model that simulates theconditions just prior to fault growth is expected to yield a fault evolution that produces greater worksavings than the cost of growing the fault. Although the work of fault propagation is small, this termis essential for predicting the timing of new fault development both in the sandbox and also in the crust.ReferencesCooke, M. L. and S. Murphy, 2004. Assessing the work budget and efficiency of fault systems usingmechanical models, JGR, 109,Del Castello, M. and M. L. Cooke, 2007. Underthrusting-accretion cycle: Work budget as revealed bythe boundary element method, JGR, 112,Cooke, M. L., M. T. Schottenfeld and S. W. Buchanan in press. Evolution of Fault Efficiency at RestrainingBends within Wet Kaolin, JSGSouloumiac, P., Maillot, B., Leroy, Y.M., 2012. Bias due to side wall friction in sandbox experiments.JSG 35, 90-101.26

Tuesday, April 16 th , 15h30-16h00On the relationship between forearc deformation, friction properties,pore pressure and megathrust rupturesNadaya Cubas 11 Caltechcubas@caltech.eduDetermining the spatio-temporal variations of frictional properties is a key issue in seismotectonics,since these properties are thought to determine the seismic potential of a subduction as well as thedeformation style of a continental margin. In this presentation, we propose to compare frictionalproperties of the regions of the 2010 Mw 8.8 Maule earthquake and the 2011 Mw 9.0 Tohoku-Okiearthquake. To retrieve these properties, two kinds of mechanical analyses are conducted. The firstone relies on the critical taper theory and yields the effective basal friction on the subduction interface,internal friction, and internal pore fluid pressure. The second is based on the limit analysis approachthat allows constraining variations of frictional properties based on the location and style of forearcfaulting. For the Maule area, we first show that the rupture area of the earthquake coincides withthe mechanically stable part of the wedge. In the surrounding area, the wedge is critical, consistentwith various evidence for active deformation. This is in particular true for the Arauco peninsula area,which seems to have stopped the Maule earthquake’s rupture to the South. This observation lendssupport to the view that the seismic rupture is inhibited when propagating beneath a critical area. Inthe frontal aseismic zone, we found a long-term hydrostatic pore pressure within the wedge and anintermediate effective friction along the megathrust (µ ∼ 0.3) probably due to the presence of clays.In the rupture area, a low effective dynamic friction (µ < 0.14) is found that probably reflects strongdynamic weakening. On the contrary, the frontal wedge of the Tohoku-Oki area is characterized bya long-term high internal pore pressure and a low effective friction along the megathrust (µ ∼ 0.1).Moreover, the earthquake activated a landward normal fault downdip of the patch of maximum slip.From the modelling of this splay fault with limit analysis, we show that the frontal wedge was submittedto a strong increase of pore pressure during the earthquake. The difference of properties of the frontalwedge of these two regions might actually reflect differences in permeability. A lower permeabilitywould enhance dynamic weakening and allow a frontal propagation of the rupture. As a consequence,forearc normal faults could then be considered as evidence for very efficient dynamic weakening alongthe Megathrust updip of their location and typify Megathrust with high tsunamigenic potential.27

Tuesday, April 16 th , 16h30-17h00Hydromechanical investigation on a CO2-like altered carbonate:a multiscale and 3D DIC approachLouis Zinsmeister 1 , 2, Jérémie Dautriat 3 , Alexandre Dimanov 2 , Michel Bornert 41 IFPEN2 Ecole Polytechnique3 CSIRO4 Institut Navierlouis.zinsmeister@IFPEN.frIn order to provide mechanical constitutive laws for reservoir monitoring during CO 2 long term storage,we studied the mechanical properties of Lavoux limestone before and after a homogeneous alterationfollowing the protocol of acid treatments defined by Egermann et al, (2006). The mechanicaldata have been analyzed at the light of systematic microstructural investigations. Firstly, the alterationimpact on the evolution of flow properties was studied at successive levels of alteration by classicalpetrophysical measurements of porosity and permeability and by observations of microstructures onthin sections and by SEM. Secondly, the mechanical properties of the samples were investigatedby classical (macro-scopic) triaxial and uniaxial tests. The study is further complemented by 2D fullfield measurements, thanks to Digital Image Correlation (DIC) performed on images acquired duringthe uniaxial tests. This technique was applied on both intact and altered materials and at differentscales of observation: (i) cm-sized samples were compressed in a classical load frame and opticallyimaged, (ii) mm-sized samples were loaded with a miniaturized compression rig implemented withina Scanning Electron Microscope. At last, 3D full field measurements were performed by 3D-DIC onmm-sized samples, which were compressed "in-situ" in X-ray microtomograph thanks to a miniaturizedtriaxial cell allowing confining pressures of up to 15 MPa. At the macroscale and for the intactsamples, a diffuse accommodation of the deformation is observed during the pseudo elastic regime,followed by sudden failure propagation after the peak stress. Conversely, the altered samples exhibitmuch more localized and pronounced deformation levels. At the SEM scale for the intact samples,but closure of the porosity, failure precursors and localized deformation were not observed. In opposition,the altered samples showed early opening of microcracks at the grain junctions. Finally, at thetomograph resolution (5 µm/voxel) and in triaxial conditions, we observed for intact samples at macroand micro scales similar behaviour as for optical and SEM characterization. At 5 MPa of confiningpressure the altered samples avoid brittle failure and a few shear bands are visible. The DIC resultssuggest that besides the overall increase of porosity, the dissolution processes enhance the localheterogeneities of the porous network, which phenomenon further increases the weakening of thematerials.ReferenceEgermann, P., Bemer, E, Zinszner, B. (2007), An experimental investigation of the rock propertiesevolution associated to different levels of CO2 injection like alteration processes, Paper SCA 2006-34, Proceeding of the international Symposium of the Society of core Analysts, Trondheim, Norway,p. 12-16.28

Tuesday, April 16 th , 17h00-17h30Multiscale laboratory and numerical rock physics - a snapshot ofongoing ROCKETH-scienceMarcel Frehner 1 , Beatriz Quintal 1 , Claudio Madonna 1 , Nicola Tisato 1 , Erik H. Saenger 11 The Rock Physics Network at ETH Zurich (ROCKETH), Switzerlamarcel.frehner@erdw.ethz.chProviding a link between rock/pore fluid properties and seismic attributes is the main goal of rockphysics and key for characterizing a subsurface fluid reservoir from seismic data. Identifying andunderstanding the physical processes in a reservoir rock at different scales is the first step and thesubject of our research. We (Quintal et al., 2011) show that combining laboratory and numericalexperiments is a powerful tool to achieve an unbiased comprehension of rock physical processes.While in laboratory experiments it is very difficult to control all the physical processes, in numericalexperiments all physical parameters can be controlled exactly. Numerically, it is even possible tostudy different physical processes separately from each other, which otherwise coexist in nature or inthe laboratory.A knowledge-feedback between laboratory and numerical studies is demonstrated on two examplesof current rock physics challenges: (i) understanding the influence of the rock microstructureon effective elastic properties and (ii) identifying the dominant physical mechanism responsiblefor intrinsic attenuation in saturated rocks at seismic frequencies. Both presented challenges aresubject to ongoing research conducted in The Rock Physics Network at ETH Zurich (ROCKETH;www.rockphysics.ethz.ch).In addition, the latest ultra-high-resolution 3D digital images of rock microstructures, obtained bySynchrotron radiation X-ray tomographic microscopy, are presented (Madonna et al., 2013). We describethis method and, to demonstrate its wide applicability, present 3D images of very different rocktypes: sandstones, dolomite, and three-phase magmas. For some samples, full and partial saturationscenarios are considered using oil, water, and air. The rock images precisely reveal the 3D rock microstructure,the pore space morphology, and the interfaces between fluids saturating the same pore.We also discuss and suggest possible applications and research directions that can be pursued onthe basis of our data.ReferencesMadonna, C., Quintal, B., Frehner, M., Almqvist, B.S.G., Tisato, N., Pistone, M., Marone, F., Saenger,E.H. (2013), Synchrotron-based X-ray tomographic microscopy for rock microstructure investigations,Geophysics, 78, D53-D64, doi:10.1190/GEO2012-0113.1.Quintal, B., Frehner, M., Madonna, C., Tisato, N., Kuteynikova, M., Saenger. E.H. (2011), Integratednumerical and laboratory rock physics applied to seismic characterization of reservoir rocks, TheLeading Edge, 30, 1360-1367, doi:10.1190/1.3672480.29

Tuesday, April 16 th , 17h30-18h00Cohesive Critical Coulomb Wedges and Rankine Limit StatesFlorian K. Lehner, 11 Geography and Geology, University of Salzburg30

Poster Presentations AbstractsApril 15 th -16 th , 12h30-14h303D thrusting in frictional wedges : comparison between experimentalobservations and numerical predictions by limit analysisPauline Souloumiac 1 , Kristian Krabbenhoft 2 , Bertrand Maillot 1 , Yves Leroy 31 Département Géosciences et Environnement, Université de Cergy-Pontoise, France2 Centre of Geotechnical Science and Engineering, Newcastle, Australia3 Laboratoire de Géologie, Ecole Normale Supérieure, CNRS, Paris, Francepauline.souloumiac@u-cergy.frThe objective is to capture 3D failure mechanisms, i.e., thrust surfaces in compressive accretionarywedges by the sole knowledge of the material and interface strengths. A 3D simple prototype composedby a flat sand layer topped with a wedge on one side and maintained between two lateralrigid walls is shortened by the relative movement of the two end walls. We consider laboratory datathat demonstrate the important influence of the side wall friction on the position of thrust faults andon spurious departures from plane strain conditions, by comparing outcomes using the two possibleconfigurations of sand boxes (fixed or mobile base) (Souloumiac et al., J. Struc. Geol., 2012). Theproposed method to reproduce these observations is the limit analysis. The 3D virtual velocity fieldis constructed by spatial discretisation. The numerical tool based on the limit analysis allows us totake into account the influence of the side wall friction and to find the 3D failure modes observed inthe laboratory at the onset of thrusting. The comparison with the analogue experiments shows theconnection between the virtual 3D velocity field and the actual deformation, and validates the numericalmethod. It is further applied to cases with lateral variations of surface slopes and basal friction toillustrate respectively the lateral transition from super to sub-critical conditions and the surface effectof patches of low friction on the décollement.ReferenceSouloumiac, P. (2012), Bias due to side wall friction in sand box experiments, Journal of StructuralGeology, 35, 90-101.31

April 15 th -16 th , 12h30-14h30Towards mechanically balanced cross-sections of the Mont TerriAnticline (Jura, Switzerland)T. Caer 1 , G. Laigneau 2 , P. Leturmy 1 , D. Frizon de Lamotte 1 , Y. Leroy 3 , B. Maillot 1 , P. Souloumiac 1 , C.Nussbaum 41 Dṕartement Géosciences et Environnement, Université de Cergy-Pontoise, France2 Ecole Centrale de Paris, France2 Laboratoire de Géologie, Ecole Normale Supérieure, CNRS, Paris, France4 Swiss Topo, Switzerlandtyphaine.caer@gmail.comThe Mont Terri Anticline is a fault-bend fold at the front of the Jura fold belt in Saint-Ursanne, Swiss.It is crossed by a highway tunnel and an underground international laboratory has been developedto study the opalinus clay layer, a rock thought to be used as repository sites for nuclear wasteelsewhere in Switzerland. Dips and layer limits are very well known along the 4 km tunnel, in thelaboratory galleries and in wells around. However, the structure of the anticline is rather badly knowncompared to the precision of the other observations, and it does not integrate all available data. The2D cross-section is challenging, with at least two kinematic phases marked by nearby competingramps and overturned beds in the back limb. In addition, the fold is evolving as double anticlines inboth NE and SW directions, exhibiting a strongly 3D structure. We propose series of cross-sectionswith several interpretations, all kinematically balanced and integrating all data. The choice betweenthese interpretations can be aided by a mechanical analysis. A preliminary analysis reveals theessential role of erosion in the force balance and therefore also in the lifetime and positions of theramps. Therefore, the first problem addressed is to identify the evolution of the topography of the foldduring its growth, precisely by force balance.32

April 15 th -16 th , 12h30-14h30Effective elastic properties of cracked solids: an experimentalcheckCéline Mallet 1 , Jérome Fortin 1 and Yves Guéguen 11 Laboratoire de Géologie, Ecole Normale Supérieure, CNRS, Paris, Francemallet@geologie.ens.frWe investigate how cracks develop and propagate in synthetic glass samples. Cracks are introducedin glass by a thermal shock of 300 o C. Crack network is documented from optical and electronic microscopyon these samples that have been submitted to a thermal shock only.Samples are cylinder of 80 mm length and 40 mm diameter. Sections were cut along the cylinderaxis and perpendicular to it. Using SEM, crack lengths and apertures can be measured. Opticalmicroscopy allows to get the crack distribution over the entire sample. The sample average cracklength is 3 mm. The average aperture is 6 +/-3 mm. There is however a clear difference between thesample core, where the crack network has approximatively a transverse isotrope symmetry and theouter ring, where cracks are smaller and more numerous.By measuring before and after the thermal treatment the radial P and S wave velocities in room conditions,and using the Effective Medium Theory, we obtained the horizontal and vertical crack density: α 1 = 0.038 and α 3 = 0.0037.Micorstructural observations allow us to obtain the oriented crack density that agree with the previousone (within a small incertitude range).These results show, for the first time, a direct confirmation of the EMT method. At the same time,they show that elastic wave velocity measurements in rocks can provide very reliable informations ontheir crack density and network.ReferenceGuéguen Y. & Kachanov M. (2011), Effective Elastic properties of Cracked Rocks - an Overview, inMechanics and crustal rocks, CISM Courses and Lectures.33

April 15 th -16 th , 12h30-14h30Deterministic chaos in frictional wedgesBaptiste Mary 1 , Bertrand Maillot 1 , and Yves M. Leroy 21 Dṕartement Géosciences et Environnement, Université de Cergy-Pontoise, France2 Laboratoire de Géologie, Ecole Normale Supérieure, CNRS, Paris, Francebaptiste.mary@u-cergy.frA triangular wedge, composed of a frictional material such as sand, and accreting additional materialat its front, is the classical prototype for accretionary wedges and fold-and-thrust belts. The SequentialLimit Analysis method is applied to capture the internal deformation to these structures resulting froma large number of faulting events during compression. The method combines the application of thekinematic approach of limit analysis to predict the optimum thrust-fold and a set of geometrical rulesto update the geometry accordingly, at each increment of shortening. It is shown that the topographyremains planar to first order with an average slope predicted by the critical Coulomb wedge theory.Failure by faulting occurs anywhere within the wedge at criticality and its exact position is sensitive totopographic perturbations resulting from the deformation history. The convergence analysis in termsof the shortening increments and of the topography discretisation reveals that the timing and theposition of a single faulting event cannot be predicted. The convergence is achieved nevertheless interms of the statistics of the distribution of the faulting events throughout the structure and during theentire deformation history. These two convergence properties plus the perturbation sensitivity justifythe claim that these compressed frictional wedges are imperfection sensitive, chaotic systems. Thisfundamental system has to be understood before considering the influence of softening on activatedramps and of erosion which are also discussed.34

April 15 th -16 th , 12h30-14h30Compressibility of low porosity carbonate rocksAurélien Nicolas 11 Laboratoire de Géologie, Ecole Normale Supérieure, CNRS, Paris, Francenicolas@geologie.ens.frThe study of the compressibility of carbonates is a major challenge due to CO2 storage projects andexploitation of oil or gas. Theoretical equations governing the compressibility of the pores, the matrixand the entire rock were developed on sandstone by Zimmerman et al. (1986), depending on thestate of stress on the medium. However, mainly due to the lack of field and experimental data, thistheory has not been validated on carbonate rocks. This study therefore focuses on the limits of theapplication of theoretical equations on carbonates. To do so, experimental data from previous studiesare compared with theoretical predictions of compressibility lower bounds (Hashin and Shtrikman,1963) to assess the adequacy of the measures. The measurements are then used in the equationscompressibility to compare the experimental to the theoretically predicted data.The problem lies in the difference between the homogeneities of sandstones and carbonates. Inaddition, limitations at the uncertainty measures are shown, which creates some uncertainty in therelevance of the formulas. Finally, the extremely limited number of complete data sets suggests thatacquisition work and new experiences are desirable.35

April 15 th -16 th , 12h30-14h30Extension failure in over-pressured frictional wedgesXiaoping Yuan 1 , Bertrant Maillot 2 , Yves M. Leroy 11 Laboratoire de Géologie, Ecole Normale Supérieure, CNRS, Paris, France2 Département Géosciences et Environnement, Université Cergy-Pontoise, Francexyuan@geologie.ens.fr, Bertrand.Maillot@u-cergy.fr, leroyy@geologie.ens.frThe objective of this research is to capture the extensive internal deformation of a frictional wedgeresting on a weak décollement using the Sequential Limit Analysis Method (SLAM) developed byCubas et al. (2008) and Mary et al. (2013). This method combines the kinematic approach oflimit analysis to determine the optimum collapse mechanism (conjugate normal faults bounding a V-shaped region moving down and pushing laterally the frontal wedge region) and simple geometricalrules to update the geometry accordingly (Xiao and Suppe, 1992).It is first shown how the kinematic approach of limit analysis for cohesionless materials reproducesexactly the critical taper theory (Dahlen, 1984) in the presence of an over-pressured décollement.Several examples are then provided to reproduce the quasi-static internal deformation of a wedge ofarbitrary initial geometry following a variation in the décollement pressure using SLAM. At each timestep, the optimum unstable collapse mechanism is determined and the wedge geometry is modifiedaccordingly. It is tentatively proposed to extend the application of this methodology to a rapidly changingdécollement pressure as it occurs typically during an Earthquake. The evolution of the structureis based then on a hierarchy of unstable collapse modes, the most unstable being first used to modifythe geometry. The possibility to validate this scheme with dedicated sand-box experiments is finallydiscussed.ReferencesCubas, N. and Leroy, Y. M. and Maillot, B. (2008), Prediction of thrusting sequences in accretionarywedges, Journal of Geophysical Research, doi: 10.1029/2008JB005717.Dahlen, F. A. (1984) Noncohesive Critical Coulomb Wedges: An Exact Solution, Journal of GeophysicalResearch, doi :10.1029/JB089iB12p10125.Mary, B.C. and Maillot B. and Leroy, Y.M. (2013), Deterministic chaos in frictional wedges revealed byconvergence analysis, Int. J. Numer. Anal. Meth. Geomech., DOI: 10.1002/nag.2177Xiao, H. And J. Suppe, 1992 (1992) Origin of rollover, AAPG Bulletin, 76, 509-525.36

April 15 th -16 th , 12h30-14h30The spatial organization of deformation in high porosity sandstones:from outcrop data to prediction of bulk fault zone propertiesElodie Saillet 1 , Zoe Shipton 2 , Christopher Wibberley 31 Institut Polytechnique LaSalle Beauvais2 Strathclyde University, Civil Engineering Department, Glasgow (UK)3 TOTAL E&P, Pointe Noire (Congo)Elodie.SAILLET@lasalle-beauvais.frUnderstanding the evolution of fault zone geometrical and hydromechanical properties during faultgrowth and network development is of major importance in fluid flow prediction in the crust. In porousrocks, faulting produces zones of deformation bands rather than planar fracture surfaces. Cataclasticdeformation bands (CDBs) are mm-cm thick brittle shear zones that form through the combined effectsof compaction and cataclasis. Porosity and grain size reduction associated with CDB formationcauses strain hardening, and the result is the initiation of a new band, adjacent to the first. Continueddeformation may possibly result in the development of localized slip planes associated to DB zones.Most commonly CDBs show a reduction of porosity, associated with a reduction of permeability. Thispermeability decrease is largest for the most evolved and thick zones of CDBs. Conversely, slipplanes can potentially have a higher permeability than the host rock. Therefore CDBs in sandstonereservoirs can potentially retard fluid flow circulation and act as barriers to fluids, whereas slip planescould be conduits for flow. Previous studies have examined the effect of connected deformation bandsystems on flow, but have not considered the effect of "open" slip planes. To predict the effect of suchstructures on fluid flow we must consider the connectivity of the relatively low permeability CDBsand any high permeability slip planes. To characterize this connectivity we have chosen to undertakedetailed mapping of structures affecting deformed sandstones in France (Provence) and USA(Utah). The principal objectives of this mapping were 1) to quantify the geometrical interconnectivitybetween low permeability CDBs and high permeability slip planes, 2) to quantify the variability of thisconnectivity at individual sites and 3) to establish if the connectivity and variations in connectivityare controlled by the host rock properties and deformation conditions. The field and laboratory datacoming from the Provence and Utah outcrops allow us to identify two different models, correspondingto different geological factors and resulting to an evolution of the deformation process. These twomodels can be identifying as a single one analogous of the fracturing sandstone reservoir, respectivelyat low and high burial depth. On the lower burial depth - dominated by poorly cohesive sands- deformation is translate by a high number of CDBs and a small amount of large but simple and noanastomosed faults. On the higher burial depth - dominated by cohesive sandstone with cement -deformation is translate by large, complex and anastomosed faults, associated with CDBs. In suchconditions, complex and high burial depth fault zones associated to cross-cutting slip surfaces canhave a strong impact on fluid flow circulation at the reservoir scale.37

April 15 th -16 th , 12h30-14h30Frictional laws of the Westerly granite: Influence of the slidingvelocity and of the normal stress on the coseismic damageFrançois Passelegue 11 Laboratoire de Géologie, Ecole Normale Supérieure, CNRS, Paris, Francepasselegue@geologie.ens.frWith the advent of high-velocity shear apparatus, several experimental studies have been conductedin recent years, improving our understanding of the evolution of fault strength and of the coseismicdamage during seismic slip.Here, we present the results of high-velocity experiments coupled with state of the art acoustic monitoring,conducted in order to study the friction law of Westerly granite and the thermo-mechanicalbehavior of rocks during large slip events. Experiments have been conducted for normal stressesand sliding velocities ranging from 5 to 20 MPa and 0.003 to 3 m/s.The increase of the normal stress and of the sliding Velocity induce, either individually or combined,(i) the decrease of the slip weakening distance (Dc) and (ii) the decrease of the residual strength ofthe fault for identical final displacements. The average values of fracture energy are independent ofthe normal stress and of the sliding velocity, while more damage are induced during experiments at0.3 m/s. By using theoretical approach, we demonstrated that most of the damage induced duringexperiments could be due to the migration of the heat front. We introduce a new fracture energysink in the seismic energy balance which can be consider as thermal cracking induced by the frictionalheating during the seismic slip. Thermal cracking could play an important role in the physicalproperties of the slip zone during earthquakes, particularly on the permeability.38