Éditeur(s) : HAL CCSD Elsevier
Résumé : International audience
ISSN: 0264-8172

hal-01664643
https://hal.archives-ouvertes.fr/hal-01664643
DOI : 10.1016/j.marpetgeo.2017.08.021

Éditeur(s) : HAL CCSD Société géologique de France
Résumé : New-Caledonia island consists of an ultramafic nappe thrusted over a continental and arc-derived basement as the result of the closure of a back-arc basin during the upper Eocene. The morphology of New-Caledonia is mostly characterized by its elongated shape and rectilinear coastline. Indeed, the western and eastern edges of the island are delineated by N140 trending normal faults that are well imaged on seismic lines. Major tectonic lineaments onland, including the scarps corresponding to the main boundary of the ultramafic nappe, also trend N140 suggesting a morphotectonic control by faults having similar kinematics (longitudinal flexure-fault of previous authors). We provide detailed observations from some of these scarps demonstrating indeed that these are not erosional features in origin. We explore the possibility that such surfaces do represent major tectonic limits that accommodated extension and significant tectonic thinning of both the peridotite nappe and its basement. Remnants of a complex extensional fault zone, 200 m wide, are still preserved along the western scarp of the Mont Dore mountain, close to Noumea city. The deformed zone is composed of an early shallow dipping detachment fault, some decimeters thick, offset by a group of late high angle normal faults and is bounded by cataclastic breccias that progressively pass into the fractured host peridotite. The late high angle fault zones have been the site of important syn-tectonic fluid circulation and are underlined by cm-thick silica infills bearing vertical striae. Typical morphology of the main scarps with break of slope is related to the combination of high angle normal faults and low angle detachment. Similar shallow dipping detachment surfaces have been observed along the east coast of southern New-Caledonia as well as within the ultramafic nappe itself and its basement. This study confirms that N140 oriented, large fault zones, with pure-normal to transtensional component of displacement, controlled the morphotectonic evolution of New-Caledonia at the regional scale, after the obduction of ophiolite nappe at the end of the Eocene. Such a fact has to be taken into account when attempting to reconstruct the post-obduction evolution of the island and more particularly when considering the development and distribution of the Ni-rich lateritic surfaces.
EISSN: 1777-5817

hal-00411855
https://hal.archives-ouvertes.fr/hal-00411855
DOI : 10.2113/gssgfbull.179.3.315

Éditeur(s) : HAL CCSD Elsevier
Résumé : International audience
The South East Sayan area,Wof the Lake Baikal is subjected to a very complex tectonic setting where the extensional stress field of the Baikal Rift System meets the compressional stress field generated by the India­Asia collision further south. Using satellite images, aerial photographs, SRTM DEM, field mapping of geomorphological structures, and published neotectonics and geological data we show that most of the relief in the SE Sayan initiated during Late Pliocene­Pleistocene through compressive reactivation of inherited structures. By Late Quaternary, clockwise rotation of the compressive field generated strike­slip faulting and local, secondary extension still within a general compressional stress field. We demonstrate that the formation of the small-scale extensional basins within the East Sayan range is not linked to general the extension in the Baikal Rift System nor to a possible asthenospheric plume acting at the base of the crust but rather to the rotation of small rigid tectonic blocks driven by the compression.
ISSN: 0743-9547

insu-00605062
https://hal-insu.archives-ouvertes.fr/insu-00605062
DOI : 10.1016/j.jseaes.2010.10.011

Éditeur(s) : HAL CCSD
Résumé : International audience
Hundreds of Paleoproterozoic gold occurrences are known in the Tapajós Gold Province, northern Brazil. However, few deposits have been analysed in detail to understand the genesis of these systems. In this study, we present data concerning the largest deposit presently known in this Province, the Tocantinzinho deposit. A strike-slip fault system controls the emplacement of the mineralized rocks and the development of the brittle fracturing that host the main gold mineralization. Hydrothermal activity played an important role in the mineralization processes.
Mineral Resources in a Sustainable World
Nancy, France

hal-01356009
https://hal.archives-ouvertes.fr/hal-01356009

Éditeur(s) : HAL CCSD Springer Verlag
Résumé : International audience
Geochronological, mineralogical, and geochemical analyses have been focussed on the Mont-Loze're- Borne plutonic complex and surrounding rocks (Ce'vennes, French Massif Central) in which B-W-Sn and As-Au-Sb mineralization is encountered. Two main results are highlighted: (1) the existence of a 301-306 Ma magmatohydrothermal event unrelated to the emplacement of the Pont-de-Montvert-Borne plutonic body at 316 Ma; (2) the magmatic and hydrothermal features are strongly associated, both in time and in space, thus demonstrating an intimate connection between mineralizing processes and magmatism in this part of the French Massif Central. We also show that mineralization and associated hydrothermal occurrences do not correspond to a simple and single geochemical signature and that a contamination model must be invoked in order to account for the complexity of isotopic results. This study demonstrates that the application of the O and H isotopic signatures as tracers of the source and nature of fluids in an orogenic context requires some specific care. Finally, a model of the tectonic-magmatic- hydrothermal evolution of the study area is suggested in which we discuss two alternative scenarios. The first one implies the existence of two different hydrothermal/ mineralizing events (Bo-W-Sn and As-Au-Sb ones). The second one suggests the same source for all hydrothermal and mineralized structures.
ISSN: 1437-3254

insu-00576546
https://hal-insu.archives-ouvertes.fr/insu-00576546
https://hal-insu.archives-ouvertes.fr/insu-00576546/document
https://hal-insu.archives-ouvertes.fr/insu-00576546/file/Chauvet-InternationalJournalEarthSciences-2012.pdf
DOI : 10.1007/s00531-011-0639-1

Éditeur(s) : HAL CCSD
Résumé : International audience
Understanding subduction rheology in both space and time has been a challenge since the advent of plate tectonics.We herein focus on "subduction infancy", which corresponds to the first 0-2 My immediately following subductionnucleation, when a newly born slab penetrates into the overriding plate mantle and heats up.The only remnants of this critical, yet elusive, geodynamic step are thin metamorphic soles, commonly found beneathpristine, 100-1000 km long portions of oceanic lithosphere emplaced on top of continents (i.e. ophiolites).In this study, we show how, during subduction infancy, transient mechanical properties of both the mantle and crustacross the subduction plate interface (during 100s ky) control and hinder the penetration of tectonic plates intothe mantle, and how this results in strong peaks of resistance and even slicing of their surface — leaving behindthin, chopped-off metamorphic slivers (i.e. metamorphic soles).These findings constrain the mechanical behaviour of the subduction plate interface (with implications for couplingprocesses and earthquake generation) as well as the properties of the crust and mantle. They also highlight the roleof fluids in enabling subduction to overcome this early resistance.
Geophysical Research Abstracts
Vienne, Austria

insu-01309386
https://hal-insu.archives-ouvertes.fr/insu-01309386

Éditeur(s) : HAL CCSD American Geophysical Union
Résumé : International audience
Phyllosilicate content and related permeability of fault zones form primary controls on hydraulic and mechanical behavior of the brittle crust. Hence, understanding and predicting the localization of these ubiquitous minerals is a major issue for fundamental and applied geosciences. We describe normal fault zones cutting a foreland arkosic turbiditic formation suffering high-T diagenesis and formed under conditions (~200°C) typical of deeply buried reservoirs and common within the seismogenic interval. Microstructural analyses show a large proportion of phyllosilicates (up to 34%) in the fault rock, derived from near-complete feldspar alteration and disaggregation during deformation. This study shows that even faults with offsets (~20 cm) much lower than bed thickness can have such large feldspar-to-phyllosilicate transformation ratios, implying that the origin of the phyllosilicates is purely transformation related. These results imply that the potential sealing capacity and strength of faults could be predicted from the host rock feldspar content. Where sealing capacity and fault strength can be related to phyllosilicate content, these properties can then also be inferred from the predicted phyllosilicate content: this opens up new horizons concerning the hydraulic and the mechanical behavior of the upper crust.
ISSN: 0094-8276

hal-00903574
https://hal.archives-ouvertes.fr/hal-00903574
DOI : 10.1002/grl.50829

Éditeur(s) : HAL CCSD
Résumé : International audience
The Madeira deposit (Sn, Nb, Ta) at Pitinga, Amazonia, in Brazil, is associated with the albite-enriched facies of the A-type Madeira granite (~1,820 Ma). Fluorine (cryolite), Y, REE, Li, Zr, and Th are potential by-products of the disseminated ore. We studied the xenotime from the core albite-enriched (CAG), transitional albite-enriched granite (TAG) and pegmatitic albite-enriched granite (PAG). In all these rocks, xenotime is magmatic, has high HREE contents (0.37 to 0.54 atoms per formula unit; LREE are not detected), and low U, Th and Ca content. The CAG xenotime is the richest in REE, followed by that in the PAG and TAG units. Fluorine was detected in all xenotime crystals from CAG (up to 5.10%) and from PAG (up to 1.40%) and in most crystals from TAG (up to 0.68%). Where the xenotime is richer in F, the Na and Si increase, the P, Ca, and the effectiveness of a thorite-type substitution decrease. Fluorine controls the ratio REE/Y; the richer the xenotime in F, the richer it is in REE, in particular Er and Yb. The cell parameters a and c are significantly shortened in F-rich xenotime. There is no evidence of OH in the structure. We contend that fluorine substitutes for O to form PO3F tetrahedra, as in bobdownsite. The formation of the PO3F tetrahedra in the xenotime accounts well for the compatibility between the crystallographic characteristics and the chemical composition.
The Canadian Mineralogist

insu-00805907
https://hal-insu.archives-ouvertes.fr/insu-00805907
DOI : 10.3749/canmin.50.6.1453

Éditeur(s) : HAL CCSD Elsevier
Résumé : International audience
The Baikal Rift System in southern Siberia is one of the main intracontinental extensional features on Earth. The rift system represents the northwestern boundary of the Amuria plate and in that respect can be considered as an evolving plate boundary. The Baikal Rift System has been widely studied both in terms of geology and geophysics and many models have been proposed for its formation and evolution. However, the age of the initiation of deformation and the mechanism driving this deformation are still largely debated. While major extension has occurred since the Late Miocene-Pliocene, the onset of extension seems older than the India-Asia collision, implying that several driving mechanisms may have acted together or in relay through time. In this work, we review the available data and models for deformation in an area encompassing the Baikal Rift System, the Sayan ranges to the west and the Transbaikal to the east. Using a synthesis of this data and our own field and mapping observations, we show that the Baikal Rift System, along with transpressional deformation in the Sayan ranges and transtension in the Transbaikal area, can be explained through major left-lateral strike-slip systems. The deformation is strongly controlled by inherited crustal and lithospheric structures, and is distributed over a wide area within the western Amuria plate that consequently cannot be considered as a rigid block. Such distributed deformation is likely to have a strong effect on the structure of the future continental margin if extension evolves towards the formation of oceanic crust.
ISSN: 1342-937X

hal-00903612
https://hal.archives-ouvertes.fr/hal-00903612
DOI : 10.1016/j.gr.2012.07.017

Éditeur(s) : HAL CCSD
Russian Academy of Science International geological meeting
Irkoutsk, Russia

insu-00610115
https://hal-insu.archives-ouvertes.fr/insu-00610115

Éditeur(s) : HAL CCSD
Résumé : International audience
Understanding subduction rheology in both space and time has been a challenge since the advent of plate tectonics. We herein focus on "subduction infancy", that is the first ~1-5 My immediately following subduction nucleation, when a newly born slab penetrates into the upper plate mantle and heats up. The only remnants of this critical yet elusive geodynamic step are thin metamorphic soles, commonly found beneath pristine, 100-1000 km long portions of oceanic lithosphere emplaced on continents (i.e., ophiolites).Through the (i) worldwide compilation of pressure-temperature conditions of metamorphic sole formation augmented by pseudosection thermodynamic modeling, (ii) calculations of the viscosity of materials along the plate interface and (iii) generic numerical thermal models, we provide a conceptual model of dynamic plate interface processes during subduction infancy (and initiation s.l.).We show in particular how major rheological switches across the subduction interface control slab penetration, and the formation of metamorphic soles. Due to the downward progression of hydration and weakening of the mantle wedge with cooling, the lower plate (basalt, sediment) and the upper plate (mantle wedge) rheologies equalize and switch over a restricted temperature-time-depth interval (e.g., at ~800°C and ~1 GPa, during 0.1-2 My, for high-temperature metamorphic sole formation). These switches result in episodes of maximum interplate mechanical coupling, thereby slicing the top of the slab and welding pieces (basalt, sediment) to the base of the mantle wedge. Similar mechanical processes likely apply for the later, deeper accretion and exhumation of high-temperature oceanic eclogites in serpentinite mélanges, or for the accretion of larger tectonic slices.This model provides constraints on the effective rheologies of the crust and mantle and general understanding, at both rock and plate scale, for accretion processes and early slab dynamics.
AGU Fall Meeting 2015
San Francisco, United States

insu-01242103
https://hal-insu.archives-ouvertes.fr/insu-01242103

Éditeur(s) : HAL CCSD
Résumé : International audience
Tin mineralization in Achmmach deposit (NE of Moroccan Central Massif) is associated with tourmaline-rich alteration halos and faults, within N70 shear zones, hosted in sandstones and metapelites of the Upper Visean-Namurian. Such deposits were reported as late-Variscan in age and related to the emplacement of late orogenic granite not outcropping in the studied area.From field constraints, it appears that the late Variscan phase is marked by a transition from transpression to extension with conditions of deformation evolving from ductile to brittle environments. During the transpressive phase (horizontal shortening direction roughly trending E-W), pervasive tourmaline is deposited under the form of metasomatic halos around fractures in the upper levels relative to the deep batholith. This hydrothermal activity is coeval with ductile shear zones within a large corridors trending N70. Tourmaline-rich alteration appears as a hardening stage contributing to the modification of the tectonic style of rocks. Thereafter, a strong brecciation occurs in response of re-using and re-opening of the strike-slip structure. The result is (1) the formation of dextral (N070, N020) and sinistral (N120) shears, and (2) the circulation of fluids within early tourmalinites and formation of tourmaline-rich breccia. Subsequently, during the late orogenic extensional phase, tourmalinite breccias reopened as normal faults acting as overpressure fluid pathways for developing other breccias and veins. Such event is marked by deposition of tin mineralization i.e., cassiterite followed by sulphides, and then fluorite and carbonates at the end.From analysis of the shear-zone structure and superimposition of tourmalines and tin mineralization in the same structures, it is possible to suggest a structural control continuum. Such control is well expressed during metasomatism and deposition of hydrothermal tourmaline under the transpressive phase. Thereafter, under the extensive phase allowing open-space, tin-mineralization were trapped.This study has been supported by AI MA/09/210
24ème RST
Pau, France

insu-01546927
https://hal-insu.archives-ouvertes.fr/insu-01546927

Éditeur(s) : HAL CCSD
Résumé : International audience
The Transbaikal region extends over several hundreds of kilometres east of the Baikal Rift System. It is characterized by a number of sub-parallel Mesozoic grabens or half grabens generally filled with late Jurassic to Early Cretaceous clastic sediments interbedded with coal layers (1). Similar basins occur on an even larger area spanning from the Transbaikal region down to Korea implying a large-scale extensional process affecting most of the Amuria plate during the Mesozoic. In the Transbaikal region, the normal faults controlling the edges of the Mesozoic basins are generally superimposed to Palaeozoic ductile shear zones implying a strong localisation of the extensional deformation on inherited structures. Recent studies, associated to our own fieldwork demonstrated that some of the faults were again activated (2), still as extensional faults, during the Tertiary or Quaternary, and that some of them are presently active. The closure of the Mongol-Okhotsk ocean separating the Siberian plate from the Amurian block during the Mesozoic corresponds to a major event in the growth process of the East Asian continent. The oceanic suture zone is situated on the southern edge of the Transbaikal region and its roughly SW-NE direction is parallel to the basins (3). The timing of the closure of the Mongol-Okhotsk ocean is still highly debated: while sedimentological and tectonic data suggest that the oceanic closure and the following collision occurred in early Middle Jurassic (4), paleomagnetic studies advocate for a Early Cretaceous collision (5). Furthermore, several other questions remain on the localization, the size and the fate of the relief that most probably formed during the collision between the Amuria block and the Siberian craton. In order to answer those questions we used low temperature thermochronology data associated to tectonic, sedimentology and palinology to investigate the evolution of the Transbaikal grabens from Mesozoic to Present. Tectonic and thermochronology data provide evidences of exhumation and erosion along the eastern edge of the Siberian craton during the Middle Jurassic as well as a potential continuum of deformation between the Mesozoic extension and the initiation of the Baikal Rift System (6). Sedimentology and palinology reveals that the sediments deposited in the Transbaikal basin did not registered large-scale compressive deformation during or after their Late Jurassic - Early Cretaceous deposition and that they do not correspond to the dismantling of a strong compressive relief.
Geophysical Research Abstracts
Vienne, Austria

insu-00810487
https://hal-insu.archives-ouvertes.fr/insu-00810487

Éditeur(s) : HAL CCSD
Résumé : International audience
The history of the peneplain in East Sayan was studied using apatite fission-track analysis (AFTA). This method is suitable for determining the formation time of the erosional surface and estimating its denudation rate. The largest known relic of the peneplanation surface in this area is the Oka Plateau, separated from the Kropotkin Ridge by the Oka-Jombolok fault. The AFTA shows that the peneplain on the Oka Plateau formed in the Late Jurassic-Early Cretaceous. This peneplain is much younger than the erosional surfaces that persist today in the Tien Shan, Gobi Altai, and Mongolian Altai (Early Jurassic). However, it is older than the peneplain on the Chulyshman Plateau, Altai (Late Cretaceous), suggesting asynchronous formation of the ancient peneplain in Central Asia. The similar exhumation histories of samples from the Oka Plateau and Kropotkin Ridge indicate that these morphotectonic structures developed from Jurassic to late Miocene as a single block, which underwent continuous slow denudation at an average rate of 0.0175 mm/yr. Active tectonic processes in the Late Miocene caused the destruction of the peneplanation surface and its partial uplifting to different altitudes. The rate of Pliocene-Quaternary vertical movements along the Oka-Jombolok fault is roughly estimated at 0.046-0.080 mm/yr, which is several times higher than the denudation rate in this area. During the Pliocene-Quaternary, the Oka Plateau has not undergone any significant morphologic changes owing to its intermediate position between the summit plain and datum surface of East Sayan and to its partial shielding by basaltic lavas.
Russian Geology and Geophysics

insu-00839215
https://hal-insu.archives-ouvertes.fr/insu-00839215
DOI : 10.1016/j.rgg.2013.06.004

Éditeur(s) : HAL CCSD Elsevier
Résumé : International audience
Despite decades of petrological and geochemical studies, the nature and setting of obducted ophiolites remain controversial: the influence of supra-subduction zone environments on pre-existing oceanic lithosphere is yet to assess, and the processes leading to subduction/obduction initiation are still poorly constrained. Our study documents successive influx of slab-derived fluids and progressive strain localization within the upper mantle in a supra-subduction environment during the first few My of the subduction history. We focus on strongly sheared mafic amphibolites intruding peridotites near the mantle–crust transition of the New Caledonia obducted ophiolite and ~ 50 to 100 m above the basal thrust contact of the ophiolite. These m- to hm-long and several m-thick shear bands are interpreted as inherited small-scale intrusions of mafic melts, probably dikes or sills, which were derived from a moderately refractory mantle source refertilized by supra-subduction zone fluids. 40Ar/39Ar age constraints on pargasite at ca. 90 Ma suggest that they could be inherited from the former Pacific west-dipping subduction.Secondary deformation of these mafic intrusions is intimately associated to three major stages of fluid infiltration: (1) the first stage of deformation and metasomatism is marked by syn-kinematic growth of Ca-amphibole (at 700–800 °C and 3–5 kbar) with a distinctive supra-subduction zone signature, and controlled later channelization of aqueous fluids. 40Ar/39Ar dating on magnesio-hornblende indicates that this deformation episode occurred at ca. 55 Ma, coincident with east-dipping subduction initiation; (2) the main metasomatic stage, characterized by the development of a phlogopite-rich matrix wrapping peridotites and amphibolite boudins, points to the percolation of alkali-rich aqueous fluids at still high temperature (650–750 °C); (3) the last, low temperature (< 600 °C) metasomatic stage results in the formation of deformed veinlets containing talc, chlorite and serpentine.
ISSN: 0024-4937

hal-01306664
https://hal.archives-ouvertes.fr/hal-01306664
https://hal.archives-ouvertes.fr/hal-01306664/document
https://hal.archives-ouvertes.fr/hal-01306664/file/Soret_2016_Strain_localization.pdf
DOI : 10.1016/j.lithos.2015.11.022

Éditeur(s) : HAL CCSD Elsevier
Résumé : International audience
We analyzed twelve fault gouge samples from the Bogd fault in south-western Mongolia to understand the origin and behavior of clay minerals. The investigation relies on x-ray powder diffraction (XRD), scanning electron microscopy (SEM), transmission electron microscopy (TEM) and high energy synchrotron x-ray diffraction methods to investigate microstructure and preferred orientation. Smectite (montmorillonite), illite-smectite mixed layers, illite-mica and kaolinite are the major clay components, in addition to quartz and feldspars, which are present in all samples. The observations suggest that the protoliths and the fault rocks were highly altered by fluids. The fluid-rock interactions allow clay minerals to form, as well as alter feldspars to precipitate kaolinite and montmorillonite. Thus, newly formed clay minerals are heterogeneously distributed in the fault zone. The decrease of montmorillonite component of some of the highly deformed samples also suggests that dehydration processes during deformation were leading to illite precipitation. Based on synchrotron x-ray diffraction data, the degree of preferred orientation of constituent clay minerals is weak, with maxima for (001) ranging from 1.3 to 2.6 multiples of a random distribution (m.r.d). Co-existing quartz and feldspars have random orientation distributions. Microstructure and texture observations of the gouges from the foliated microscopic zone, alternating with micrometric isotropic clay-rich area, also indicate that the Bogd fault experienced brittle and ductile deformation episodes. The clay minerals may contribute to a slip weakening behavior of the fault.
ISSN: 0191-8141

insu-00667194
https://hal-insu.archives-ouvertes.fr/insu-00667194
DOI : 10.1016/j.jsg.2011.10.006

Éditeur(s) : HAL CCSD
Résumé : International audience
The Achmmach tin mineralization (NE of theMoroccan Central Massif) is associated with tourmaline-richalteration halos, veins, and faults hosted in sandstones andmetapelites of the Upper Visean-Namurian. These depositsare reported to be late Hercynian in age and related to theemplacement of late-orogenic granite not outcropping in thestudied area. Structural and paragenetic studies of theAchmmach tin deposit were conducted in order to establisha general model of the mineralization. From field constraints,the late Hercynian phase is marked by a transition fromtranspression to extension with deformation conditions evolvingfrom ductile to brittle environments. The transpression(horizontal shortening direction roughly trending E-W) is coevalwith the emplacement of the first tourmaline halos alongseveral conjugated trends (N070, N020, and N120).Thereafter, a tourmaline-rich breccia formed in response tothe fracturing of early tourmaline-altered rocks.Subsequently, during the extensional phase, these structureswere reactivated as normal faults and breccias, allowing theformation of the main tin mineralization (cassiterite) associatedwith a wide variety of sulfides (arsenopyrite, chalcopyrite,sphalerite, galena, pyrrhotite, bismuthinite, pyrite, and stannite). This evolution ends with fluorite and carbonate deposition.The hydrothermal fluid flow velocity, calculated byapplying statistical measures on the tourmaline growth bands,varies with the lithology. Values are lower in metapelites andhigher in breccia. In the general evolution model proposedhere, tourmaline alteration makes the rock more competent,allowing for brittle fracturing and generation of open spacewhere the main Sn mineralization was precipitated.
Mineralium Deposita

insu-01240481
https://hal-insu.archives-ouvertes.fr/insu-01240481
DOI : 10.1007/s00126-015-0613-0

Éditeur(s) : HAL CCSD Wiley-Blackwell
Résumé : International audience
P>Most fault-related pseudotachylytes in the Val Sesia area, Ivrea zone, NW Italian Alps, occur in two areas: near the Canavese Line in the Kinzigite Formation paragneisses and in an elongate belt in the Mafic Complex gabbros away from the Canavese Line. Ar-40/Ar-39 dating indicates that pseudotachylytes from the Kinzigite Formation were formed in the late Eocene, likely during the early stages of the collision between the NW corner of the Adriatic indenter and the European margin. Pseudotachylytes from the Mafic Complex were formed under a transpressive strain regime, as suggested by kinematic indicators, and yield Early Cretaceous (Albian) Ar-40/Ar-39 ages. Their formation is possibly related to shortening during the inception of the tectonic inversion of the Adriatic passive margin.
ISSN: 0954-4879

hal-00617518
https://hal.archives-ouvertes.fr/hal-00617518
DOI : 10.1111/j.1365-3121.2011.00994.x