Éditeur(s) : HAL CCSD Springer Verlag
Résumé : We present a preliminary probabilistic tsunami hazard assessment of Canadian coastlines from local and far-field, earthquake, and large submarine landslide sources. Analyses involve published historical, palaeotsunami and palaeoseismic data, modelling, and empirical relations between fault area, earthquake magnitude, and tsunami run-up. The cumulative estimated tsunami hazard for potentially damaging run-up (≥1.5 m) of the outer Pacific coastline is ~40-80 % in 50 years, respectively one and two orders of magnitude greater than the outer Atlantic (~1-15 %) and the Arctic (<1 %). For larger run-up with significant damage potential (≥3 m), Pacific hazard is ~10-30 % in 50 years, again much larger than both the Atlantic (~1-5 %) and Arctic (<1 %). For outer Pacific coastlines, the ≥1.5 m run-up hazard is dominated by far-field subduction zones, but the probability of run-up ≥3 m is highest for local megathrust sources, particularly the Cascadia subduction zone; thrust sources further north are also significant, as illustrated by the 2012 Haida Gwaii event. For Juan de Fuca and Georgia Straits, the Cascadia megathrust dominates the hazard at both levels. Tsunami hazard on the Atlantic coastline is dominated by poorly constrained far-field subduction sources; a lesser hazard is posed by near-field continental slope failures similar to the 1929 Grand Banks event. Tsunami hazard on the Arctic coastline is poorly constrained, but is likely dominated by continental slope failures; a hypothetical earthquake source beneath the Mackenzie delta requires further study. We highlight areas susceptible to locally damaging landslide-generated tsunamis, but do not quantify the hazard.
ISSN: 0921-030X

hal-00963614
https://hal.archives-ouvertes.fr/hal-00963614
DOI : 10.1007/s11069-013-0809-5

Éditeur(s) : HAL CCSD AGU and the Geochemical Society
Résumé : International audience
The Ganos fault is the westernmost segment of the North Anatolian Fault that experienced the Mw = 7.4 earthquake of 9 August 1912. The earthquake revealed 45-km-long of surface ruptures inland, trending N70 E, and 5.5 m of maximum right lateral offset near Güzelköy. The long-term deformation of the fault is clearly expressed by several pull-apart basins and sag ponds, pressure and shutter ridges and offset streams. In parallel with detailed geomorphologic investigations, we measured co-seismic and cumulative displacements along the fault, and selected the Güzelköy site for paleoseismology. A microtopographic survey at the site yields 10.5 0.5 m and 35.4 1.5 m cumulative lateral offsets of stream channels and geomorphologic features. Seven paleoseismic parallel and cross-fault trenches document successive faulting events and provide the timing of past earthquakes on the Ganos fault segment. Radiocarbon dating of successive colluvial wedges in trench T1, and the fresh scarplet above (probably 1912 surface rupture) indicate the occurrence of three faulting events since the 14th century. Parallel trenches (3, 5, 6 and 7) expose paleo-channels and show a cumulative right-lateral offset of 16.5 1.5 m next to the fault, and 21.3 1.5 m total channel deflection. Radiocarbon dating of past channel units and fault scarp-related colluvial deposits imply an average 17 +/ 5 mm/year slip rate and 323 142 years recurrence interval of large earthquakes during the last 1000 years on the Ganos fault. The succession of past faulting events and inferred slip rate west of the Marmara Sea provide more constraint on the long-term faulting behavior in the seismic gap of the North Anatolian Fault and may contribute to a better seismic hazard assessment in the Istanbul region.
ISSN: 1525-2027

hal-00748218
https://hal.archives-ouvertes.fr/hal-00748218
DOI : 10.1029/2011GC003960

Éditeur(s) : Amer Geophysical Union
Résumé : The oblique convergence between North American and Caribbean plates is accommodated in a bookshelf faulting manner by active, oblique-normal faults in the northern part of the Lesser Antilles arc. In the last 20 years, two M > 6 earthquakes occurred along a large, arc parallel, en echelon fault system, the 16 March 1985 in Redonda and 21 November 2004 in Les Saintes. A better understanding of active faulting in this region permit us to review the location and magnitude of historical earthquakes by using a regional seismic attenuation law. Several others moderate earthquakes may have occurred along the en echelon fault system implying a strong seismic hazard along the arc. These faults control the effusion of volcanic products and some earthquakes seem to be correlated in time with volcanic unrest. Shallow earthquakes on intraplate faults induced normal stress and pressure changes around neighboring volcano and may have triggered volcanic activity. The Redonda earthquake could have initiated the 1995 eruption of Montserrat's Soufriere Hills by compressing its plumbing system. Conversely, pressure changes under the volcano increased Coulomb stress changes and brought some faults closer to failure, promoting seismicity. We also discuss the magnitude of the largest 11 January 1839 and 8 February 1843 megathrust interplate earthquakes. We calculate that they have increased the stress on some overriding intraplate faults and the extensional strain beneath several volcanoes. This may explain an increase of volcanic and seismic activity in the second half of the 19th century culminating with the devastating, 1902 Mount Pelee eruption.
Journal Of Geophysical Research-solid Earth (0148-0227) (Amer Geophysical Union), 2011-10 , Vol. 116 , N. B10308 , P. 1-26

Droits : Copyright 2011 by the American Geophysical Union
http://archimer.ifremer.fr/doc/00200/31097/29515.pdf
DOI:10.1029/2011JB008443
http://archimer.ifremer.fr/doc/00200/31097/

Éditeur(s) : HAL CCSD Elsevier
Résumé : International audience
The ∼55 km-long Firouzkuh fault is located in the Central Alborz Mountains of Iran. It is a left-lateral fault, which dips to the south, and possesses a small dip-slip component of motion that we interpret to result from extension. The ratio of horizontal to vertical displacement across the fault, calculated from the cumulative displacement of landscape features, is 7.6. We provide constraints on the timing of the last earthquake on the Firouzkuh fault from two trenches (T1 and T2) across the fault zone, excavated in 2004, and located east of Firouzkuh city. The trenches expose faulted sedimentary deposits. Two optically-stimulated luminescence (OSL) ages from sediments in the lower part of trench T1 date from the late Pleistocene (15.9 ± 0.9 ka and 27.1 ± 1.7 ka). The younger of the two dated units in T1 is displaced vertically across the fault by 2.2–4.4 m, from which we estimate a strike-slip displacement of 18.2–33.4 m, and hence a average horizontal slip-rate of 1.1–2.2 mm/yr. The sediments exposed in T1 do not yield constraints on the most recent earthquake history. In trench T2, however, human skeletal remains of a middle aged male, which yield a radiocarbon age of 1159 ± 28 BP (corresponding to a mean calendar age of 791 AD), were found within a faulted alluvial layer at a depth of 60–70 cm from the surface. The existence of these medieval human places shows that a surface-rupturing earthquake occurred at some time after 1159 ± 28 BP. The amount of slip in each earthquake on the Firouzkuh fault is difficult to estimate, but assuming the entire ∼55 km fault length ruptures in each event, they will have had a maximum magnitude of 7.1. At our estimated late Quaternary slip-rate of ∼1.1–2.2 mm/yr magnitude 7.1 earthquakes, involving ∼1.2 m average displacement, would be expected to occur every ∼1100–540 years. As the last earthquake on the Firouzkuh fault may be up to ∼700 years in age we suggest that the Firouzkuh fault is a major hazard for earthquakes in the near future.
ISSN: 0743-9547

hal-01122835
https://hal.archives-ouvertes.fr/hal-01122835
DOI : 10.1016/j.jseaes.2013.12.018

Éditeur(s) : Soc Geol France
Résumé : The Lesser Antilles subduction zone has produced no recent strong thrust earthquakes, making it difficult to quantify the seismic hazard from such events. The Lesser Antilles arc has a low subduction rate and an accretionary wedge that is very wide at its southern end. To investigate the effect of the wedge on seismogenesis, numerical models of forearc thermal structure were constructed along six transects perpendicular to the arc in order to determine the thermally predicted width of the seismogenic zone. The geometry of each section is constrained by published seismic profiles and crustal models derived from gravity and seismic data and by earthquake hypocenters at depth. A major constraint on the deep part of the model is that mantle temperature beneath the volcanic arc should achieve a temperature of 1,100 degrees C to generate partial melts. Predicted surface heat flow is compared to the available heat flow observations. Thermal modeling results indicate a systematic southward increase in the width of the seismogenic zone, more than doubling in width from north to south and corresponding to a dramatic southward increase in forearc width (distance from the arc to the deformation front of the accretionary wedge). The minimum width of the seismogenic zone (distance between the intersections of the subduction interface with the 150 degrees C and 350 degrees C isotherms) increases from about 80 km, north of 16 degrees N, to 230 km, at 13 degrees N. The maximum width (between the 100 degrees C and 450 degrees C isotherms) ranges from about 150 km in the north to up to 320 km in the south. This large variation in the width of the seismogenic zone is a consequence of the increasing width of the accretionary wedge to the south, caused by the increased thickness of sediment on the subducting plate. There is good agreement between the thermally predicted seismogenic limits and the sparse distribution of recorded thrust earthquakes, which are observed only in the northern portion of the arc. Possible scenarios for mega-thrust earthquakes are discussed. Depending on the segment length (along-strike) of the rupture plane, the occurrence of an event of magnitude 8-9 cannot be excluded.
L’absence de grands séismes récents à mécanismes chevauchants dans la zone de subduction des Petites Antilles rend difficile l’évaluation de l’aléa sismique lié à de tels événements. L’arc des Petites Antilles est caractérisé par une faible vitesse de subduction et par la présence d’un prisme d’accrétion très développé à son extrémité méridionale. Afin d’évaluer les effets de la largeur de ce prisme sur la genèse des séismes, nous avons étudié six sections perpendiculaires à l’arc, du nord au sud de celui-ci, pour déterminer la largeur de la zone sismogène prédite par les modèles thermiques appliqués à chacune de ces coupes. La géométrie de ces dernières est contrainte par les profils sismiques publiés, par les modèles de structure crustale déduits des données gravitaires et sismiques, et enfin par la distribution des hypocentres des séismes. Un contrôle important permettant de tester la validité des modèles thermiques en profondeur est qu’une température minimale de 1 100oC, compatible avec la fusion partielle du manteau hydraté, doit être atteinte sous l’arc volcanique actif. Par ailleurs, le flux thermique en surface prédit par ces modèles doit être compatible avec les mesures de flux de chaleur. Les modèles thermiques retenus d’après ces critères montrent une augmentation du simple au double vers le sud de la largeur de la zone sismogène, qui correspond à un élargissement considérable de la taille du domaine avant-arc. En effet, la largeur minimale de la zone sismogène (définie comme la distance entre les intersections de l’interface des plaques avec les isothermes 150o et 350oC) augmente d’environ 80 km au nord de 16oN jusqu’à 230 km à 13oN. La largeur maximale de cette zone (définie par les intersections de l’interface avec les isothermes 100o et 450oC) augmente, quant à elle, d’environ 150 km au nord jusqu’à 320 km au sud de l’arc. Cette variation considérable est la conséquence de l’augmentation de la largeur du prisme d’accrétion, elle-même causée par l’accumulation croissante des sédiments déposés sur la plaque plongeante. Les largeurs de la zone sismogène prédites à l’aide des modèles thermiques sont en bon accord avec les rares données disponibles sur les séismes à mécanismes chevauchants dans la partie nord de l’arc. Les scénarios possibles relatifs à des méga-séismes de ce type n’excluent pas de futurs événements atteignant des magnitudes de 8 à 9.
Bulletin De La Societe Geologique De France (0037-9409) (Soc Geol France), 2013 , Vol. 184 , N. 1-2 , P. 47-59

Droits : 2013 Societe Geologique de France
http://archimer.ifremer.fr/doc/00140/25140/29432.pdf
DOI:10.2113/gssgfbull.184.1-2.47
http://archimer.ifremer.fr/doc/00140/25140/

Éditeur(s) : HAL CCSD Oxford University Press (OUP)
Résumé : International audience
To estimate the seismic hazard, the geometry (dip, length and orientation) and the dynamics (type of displacements and amplitude) of the faults in the area of interest need to be understood. In this paper, in addition to geomorphologic observations, we present the results of two ground penetrating radar (GPR) campaigns conducted in 2010 and 2011 along the Emeelt fault in the vicinity of Ulaanbaatar, capital of Mongolia, located in an intracontinental region with low deformation rate that induces long recurrence time between large earthquakes. As the geomorphology induced by the fault activity has been highly smoothed by erosion processes since the last event, the fault location and geometry is difficult to determine precisely. However, by using GPR first, a non-destructive and fast investigation, the fault and the sedimentary deposits near the surface can be characterized and the results can be used for the choice of trench location. GPR was performed with a 50 MHz antenna over 2-D lines and with a 500 MHz antenna for pseudo-3-D surveys. The 500 MHz GPR profiles show a good consistency with the trench observations, dug next to the pseudo-3-D surveys. The 3-D 500 MHz GPR imaging of a palaeochannel crossed by the fault allowed us to estimate its lateral displacement to be about 2 m. This is consistent with a right lateral strike-slip displacement induced by an earthquake around magnitude 7 or several around magnitude 6. The 2-D 50 MHz profiles, recorded perpendicular to the fault, show a strong reflection dipping to the NE, which corresponds to the fault plane. Those profiles provided complementary information on the fault such as its location at shallow depth, its dip angle (from 23 • to 35 •) and define its lateral extension. Central Asia is known for its high level of seismic hazards, especially Mongolia, which has been one of the most seismically active intracontinental regions in the world with four large earthquakes (magnitude around 8) along its active faults in the western part of the country during the last century (Khilko et al. 1985). The deformation in Mongolia is located between compressive structures related to the collision and penetration of the Indian plate into the Eurasian plate and extensive structures in the north of the country related with the Baykal rift (Tapponnier & Molnar 1979; Baljinnyam et al. 1993; Schlupp 1996; Bayasgalan & Jackson 1999). The seismic activity observed in the vicinity of Ulaanbaatar (UB), capital of Mongolia, is relatively low compared to the activity observed in western Mongolia. Nevertheless, since 2005, the seismic activity around UB not only has increased, but is also organized (see Fig. 1) at the west of UB along two perpendicular directions, which determine two active faults: Emeelt fault, discovered in 2008 (NNW-SSE direction, 25-km-long minimum and situated about 10 km W of UB) and Hustai fault (WSW–ENE direction, 80 km long, with its NE tip at less than 20 km west of UB); their length and morphology indicate that they can produce earthquakes of magnitude 6.5–7.5 (Schlupp et al. 2012). Most of the Mongolian population (1.2 million over 3 million) is concentrated at UB, which is the main political and economical centre of the country. Hence, the study of seismic hazard and the estimation of the probability of future destructive earthquakes are of primary importance for the country (Dugarmaa et al. 2006). Since the last large earthquake, the faults geomorphology has been highly smoothed by erosional processes and the exact location of the fault plane surface rupture is thus hidden within a several metre wide strip. The GPR method has been proven to give good and useful results to characterize faults by identifying offsets of radar reflections (Malik et al. 2007; Christie et al. 2009; Yalçiner et al. 2013) and
ISSN: 0956-540X

hal-01342629
https://hal.archives-ouvertes.fr/hal-01342629
https://hal.archives-ouvertes.fr/hal-01342629/document
https://hal.archives-ouvertes.fr/hal-01342629/file/Dujardin_Bano_etal_Geoph.%20J.%20Int._2014_final.pdf
DOI : 10.1093/gji/ggu130

Éditeur(s) : Université Nice Sophia Antipolis
Résumé : The Lesser Antilles is a case study of a very slow subduction (~2 cm/yr) of an old oceanic lithosphere (~84-100 Ma). The region presents a relatively low seismic activity, especially along the interplate contact, and the seismic hazard associated with a possible mega-thrust earthquake is still poorly known. This PhD thesis is a first step toward assessing the ability of the Lesser Antilles subduction zone to produce such a large subduction event. To do so, it aims at constraining the downdip width of the interplate's seismogenic zone. The lack of coverage of permanent seismological stations is a major limitation in the exploration of the Lesser Antilles subduction zone. It is due to the presence of only small aligned islands at far distances from the potentially seismogenic interplate area. Several oceanographic cruises were therefore planned that notably allowed the repeated deployment of ocean bottom seismometers; some of them being left for a few months of background seismicity recording. This thesis specifically focuses on two sets of wide-angle seismic data acquired offshore the Dominica and Martinique islands. From their analysis 3D and 2D tomographic models were produced respectively over the forearc region and across the whole subduction complex. These models constrain the plates' seismic structure as well as their geometry. They allow the discussion of how the imaged structures affect the subduction processes and give a first estimation of the downdip width of the seimogenic zone, defined as the segment of the interplate between the backstop and the upper plate's Moho. The joint interpretation of seismic models and earthquake localizations then refine this first assessment. Epicenter distribution from height months of seismic recording shows indeed that seismicity concentrates within the inner forearc region. The strong velocity gradient that characterize its basement suggests it is denser and more rigid than the more deformable outer forearc basement. The updip limit of the seismogenic zone could then lie arcward of the backstop at the contact of the interplate and the seaward end of the inner forearc crust. At depth, interplate earthquake mechanisms are observed between 35 and 45 km depth and interpreted to have occurred at the downdip limit of the seismogenic zone. The latter could reach a depth over 10,km deeper than the contact of the upper plate's Moho with the interplate, and therefore lie within the mantle wedge. All together, these results imply a large downdip width of the seismogenic zone (~70 km) offshore the Dominica and Martinique islands. Further work is, however, needed in order to fully comprehend the ability of the Lesser Antilles subduction zone to produce a possible mega-thrust earthquake. This would necessitate the evaluation of seismic coupling at the interplate contact and the possible segmentation of the seismogenic zone, for instance, due to the subduction of oceanic ridges.
Les Petites Antilles présentent un contexte géodynamique caractérisé par la subduction à très faible vitesse (~2 cm/an) d'une lithosphère océanique âgée (~84-100 Ma). L'activité sismique y est relativement faible, notamment à l'interplaque, où l'aléa sismique lié à un éventuel séisme de méga-chevauchement reste encore mal contraint. Cette thèse se veut être une première étape dans l'évaluation de la capacité de la zone de subduction des Petites Antilles à générer un tel évènement. Dans ce but, ces travaux tentent d'appréhender l'extension en profondeur du domaine sismogène de l'interplaque. Le manque de couverture des stations sismologiques permanentes est un inconvénient majeur dans l'exploration des Petites Antilles. Il s'explique en raison du peu de terres émergées et de leur éloignement de la zone potentiellement sismogène de l’interplaque. La région a donc fait l'objet de plusieurs campagnes océanographiques qui ont permis, notamment, le déploiement de sismomètres fond de mer (OBS); certains instruments étant restés immergés plusieurs mois afin de procéder à une écoute de la sismicité. Les travaux réalisés au cours de cette thèse se sont focalisés sur deux jeux de données de sismique grand-angle acquis au large des îles de la Dominique et de la Martinique. Leur analyse a permis la construction de modèles tomographiques 3D et 2D respectivement à l'échelle de l'avant-arc et de l'ensemble de la subduction. Ces modèles renseignent sur la structure sismique des plaques en convergence ainsi que sur leur géométrie. Ils permettent de discuter le rôle de la structure dans le fonctionnement de la subduction et d’obtenir une première estimation de l'extension en profondeur de la zone sismogène en considérant la portion de l'interplaque comprise entre le butoir et le Moho de la plaque supérieure. Dans un second temps, l'interprétation conjointe des modèles tomographiques et des localisations des séismes locaux a permis d'affiner cette estimation. Huit mois d'enregistrement de la sismicité montrent en effet une concentration des épicentres dans la région interne de l’avant-arc. Celle-ci présente un socle épais, à fort gradient de vitesse interprété comme plus dense et rigide que le socle de la région externe de l'avant-arc, plus déformable. La limite amont de la zone sismogène pourrait donc se situer en retrait de la position du butoir au contact de l'interplaque et de la limite entre ces deux zones de l'avant-arc. En profondeur, des mécanismes interplaques sont observés entre 35 et 45 km et interprétés comme des marqueurs de la limite aval de la zone sismogène. Cette dernière pourrait donc atteindre une profondeur jusqu'à 10 km supérieure à la limite précédemment évoquée et se trouver, par conséquent, au contact du manteau lithosphérique. L'ensemble de ces résultats suggèrent une extension en profondeur de la zone sismogène (i.e une largeur) de près de 70 km face aux îles de la Dominique et de la Martinique. Ces travaux doivent cependant être poursuivis afin d’évaluer pleinement la capacité de la zone de subduction des Petites Antilles à générer un éventuel séisme de méga-chevauchement. Le taux de couplage à l'interplaque doit être précisé ainsi que sa possible segmentation en lien, par exemple, avec à l'entrée en subduction des rides océaniques.

Droits : info:eu-repo/semantics/openAccess
http://archimer.ifremer.fr/doc/00172/28335/26628.pdf
http://archimer.ifremer.fr/doc/00172/28335/

Éditeur(s) : HAL CCSD Elsevier
Résumé : International audience
The Gulf of Cadiz spans the plate boundary between Africa and Eurasia west of the Betic–Rif mountain belt. A narrow east dipping subduction zone descends beneath the Gulf of Cadiz and the straits of Gibraltar. The deep crustal structure of the Gulf and the adjacent SW Iberian and Moroccan margins is constrained by numerous multi-channel seismic reflection and wide-angle seismic surveys. A compilation of these existing studies is presented in the form of depth to basement, sediment thickness, depth to Moho and crustal thickness maps. These structural maps image an E–W trending trough, with thin (< 10 km) crust beneath the Gulf of Cadiz. This trough is filled by an eastward thickening wedge of sediments, reaching a thickness of 10–15 km in the eastern Gulf. These sediments are tectonically deformed, primarily along a series of westward-vergent thrust faults and represent a 200–250 km wide accretionary wedge. The northern and especially the southern limits of the accretionary wedge are marked by sharp morphological lineaments showing evidence of recent deformation. These tectonic limits are situated in an internal position with respect to the Miocene deformation front (external Betic and Rif allocthons), which has been abandoned. At the western boundary of the accretionary wedge, near the adjacent Seine and Horseshoe abyssal plains, an E–W trending basement high (Coral Patch Ridge) can be seen indenting the deformation front in an asymmetric manner. Analog modeling is performed using granular materials accreted against a semicircular backstop (representing the basement of the Rif and Betic mountain belts). The modeling initially produces a symmetric, arcuate accretionary wedge. The ensuing collision of an oblique rigid indenter retards accretion on one side, resulting in an embayment and a locally steeper deformation front. The deformation pattern observed in morphology and high-resolution seismic profiles suggests the accretionary wedge and underlying subduction system is still active. The implications of active subduction for the source region of the 1755 Lisbon earthquake and the regional seismic hazard assessment are discussed.
ISSN: 0040-1951

hal-00424922
http://hal.univ-brest.fr/hal-00424922
http://hal.univ-brest.fr/hal-00424922/document
http://hal.univ-brest.fr/hal-00424922/file/Gutscher_09Tphys_rev.pdf
DOI : 10.1016/j.tecto.2008.11.031

Éditeur(s) : HAL CCSD
Résumé : The purpose of this work is to specify the deformation offshore East of Taiwan by characterizing major active faults in particular at the level of the Ryukyu forearc. The major problem associated with the study of this region is that earthquakes are located outside the permanent networks. Consequently, this led to some unprecise earthquake locations limiting tectonic and statistical analyses, and therefore the correct estimation of the seismic hazard. This work is included in the project ANR ACTS (Active Tectonics and Seismic Hazard in Taiwan). To achieve this goal, we worked on three different timescales for which a different localization method has been proposed to obtain a clear image of seismic deformation and highlight major faults offhore East of Taiwan. First, at the scale of the century (from the analysis of a seismicity catalogue with homogeneous magnitude), the instrumental historical earthquakes (1897-2007) show that the Ryukyu margin was affected by 6 earthquakes of magnitude greater than 7 with 4 of them for which the responsible fault is unknown. We propose a method of relative location to relocate the historic events from the time difference of P and S arrival times (Ts-Tp) at each station. The location is obtained by searching for "analogue" earthquakes in the recent instrumental catalogue (1991-2008) for which Ts-Tp at each station is close. This method has been applied to the largest (in terms of magnitude) earthquake ever recorded in Taiwan, the earthquake of June 5, 1920 (M7.7 +/- 0.2). A rupture along the interplate of the Ryukyu subduction with a possible nucleation at the downdip limit of a splay-fault is certainly responsible of this earthquake. Then, at the scale of the decade, the improvement of the method of absolute location based on MAXimum of Intersection of EDT (MAXI, EDT: Equal Differential Time) allows to better extract the erroneous arrival times and prevents the "trade-off" between depth and origin time, and between depth and epicentral position when the azimuthal coverage is low (gap Azimuthal > 180degre). Synthetic tests show the effectiveness of MAXI (using P-wave only) to determine the parameters x, y and z even when the azimuthal gap is important. In the latter case, the quality of the results is dependent on the velocity model to represent the 3D structure of the Earth. We propose an approach involving the use of a 3D a priori P-wave velocity model to locate earthquakes which are lateral and remote to the seismic network. An application to the case of Taiwan and Ecuador validates this approach. Finally, at the level of a few months, the marine geophysical campaign RATS (Ryukyu Arc: Tectonics and Seismology) was conducted in two stages. A passive seismology experiment from July to October, 2008 (RATS1) has been conducted over the forearc of the Ryukyus and then active seismic experiment (refraction and reflection) was conducted in May 2009 in a NNE - SSW line through the Ryukyu margin. These two combined experiments allow improving our knowledge about the crustal structure of the margin. At the level of the forearc, the basement of the forearc is characterized by vertical backstop and a very deformed basis certainly associated with a significant out of sequence deformation. In depth, the downgoing plate is certainly affected by a tear that controls the seismicity in the region of transition between the subduction and collision.
Le but de ce travail est de préciser la déformation en mer à l'est de Taïwan et de notamment caractériser les failles actives majeures en particulier au niveau de l'avant-arc des Ryukyus. Le problème majeur associé à l'étude de cette région est que les séismes sont localisés en dehors des réseaux permanents et que cela conduit à des localisations peu précises limitant les analyses statistiques, tectoniques et donc l'estimation correcte de l'aléa sismique. Ce travail est inclus dans le projet ANR ACTS (Active Tectonics and Seismic Hazard in Taiwan). Pour atteindre cet objectif, nous travaillons à trois échelles de temps différentes pour lesquelles une méthode de localisation différente a été proposée pour obtenir une image précise de la déformation sismique et mettre en évidence les failles majeures en mer à l'est de Taïwan. D'abord, à l'échelle du siècle (à partir de l'analyse d'un catalogue de sismicité homogène en magnitude), les séismes historiques instrumentaux (1897-2007) montrent que la marge des Ryukyus a été affectée par 6 séismes de magnitude supérieure à 7 dont 4 pour lesquels la faille responsable n'est pas connue. Nous proposons une méthode de localisation relative qui permet de relocaliser les évènements historiques à partir des différences de temps d'arrivées P et S (Ts-Tp) à chaque station. La localisation est obtenue en recherchant les séismes analogues dans le catalogue instrumental récent (1991-2008) pour lesquels Ts-Tp à chaque station est proche. Cette méthode a été appliquée au plus gros séisme (en terme de magnitude) jamais enregistré à Taïwan, le séisme du 5 juin 1920 (M7.7 $pm$ 0.2). Ce séisme a certainement pour origine une rupture sur l'interplaque sismogène de la subduction des Ryukyus avec une possible nucléation à la base d'une faille hors-séquence. Ensuite, à l'échelle de la dizaine d'années, l'amélioration de la méthode de localisation absolue du MAXimum d'Intersection EDT (MAXI, EDT: Equal Differential Time) permet de mieux extraire les temps d'arrivées erronés et empêche les "trade-off" entre profondeur et temps d'origine et, entre profondeur et position épicentrale lorsque la couverture azimutale est faible (gap azimutal > 180°). Des tests synthétiques montrent l'efficacité de MAXI (en utilisant les ondes P uniquement) à déterminer les paramètres x, y et z même lorsque le gap azimutal est important. Dans ce dernier cas, la qualité des résultats reste dépendant du modèle de vitesse à représenter la structure 3D de la Terre. Nous proposons une démarche associant l'utilisation d'un modèle de vitesse des ondes P 3D a priori avec la méthode MAXI pour localiser les séismes qui sont latéraux et éloignés du réseau. Une application au cas de Taïwan et en Équateur valide cette approche. Enfin, à l'échelle de quelques mois, la campagne de géophysique marine RATS (Ryukyu Arc: Tectonics and Seismology) a été menée en deux temps. Une expérience de sismologie passive de juillet à octobre 2008 (RATS1) a été menée au-dessus de l'avant-arc des Ryukyus puis une expérience de sismique active (réfraction et réflexion) a été menée en mai 2009 sur une ligne NNE-SSW à travers la marge des Ryukyus. Ces deux expériences combinées permettent d'améliorer notre connaissance de la structure crustale de la marge. Au niveau de l'avant-arc, le socle de l'avant-arc est caractérisé par une extrémité formant un butoir vertical et une base très déformée certainement associée à une importante déformation hors-séquence. En profondeur, la plaque plongeante est certainement affectée par une déchirure qui contrôle la sismicité dans cette région de transition entre la subduction et la collision.
https://tel.archives-ouvertes.fr/tel-00660737

tel-00660737
https://tel.archives-ouvertes.fr/tel-00660737
https://tel.archives-ouvertes.fr/tel-00660737/document
https://tel.archives-ouvertes.fr/tel-00660737/file/Thesis_small.pdf

Éditeur(s) : HAL CCSD Elsevier
Résumé : International audience
We use a strike-slip fault analog model to study experimentally the role played by along-fault non-uniform and asymmetric applied normal stress on both coseismic slip and long-term fault behavior. Our model is based on a visco-elasto-plastic multi-layered rheology that allows to produce several hundreds of scaled analog microquakes and associated seismic cycles. Uniform or heterogeneous applied normal stress along the fault plane is imposed and maintained constant during the whole experiment durations. Our results suggest that coseismic slip patterns are strongly controlled by spatial normal stress variations and subsequent accumulated shear stress along fault strike. Major microquakes occur preferentially in zones of major shear stress asperities. Coseismic slip distributions exhibit a pattern similar to the along-fault applied normal stress distribution. The occurrence of isolated low to moderate microquakes where residual stresses persist around secondary stress asperities, indicates that stress conditions along the fault also control the whole variability of fault slip events. Moreover, when fault slip stability conditions are modulated by normal stress distribution, our experiments suggest that the along-fault stress heterogeneity influences the seismic cycle regularity and, consequently, long-term fault slip behavior. Uniform applied normal stress favors irregular seismic cycles and the occurrence of earthquakes clustering, whereas non-uniform normal stress with a single high amplitude stress asperity generates strong characteristic microquake events with stable return periods. Together our results strengthen the assumption that coseismic slip distribution and earthquake variability along an active fault may provide relevant information on long term tectonic stress and could thus improve seismic hazard assessment.
ISSN: 0012-821X

hal-01666379
https://hal.archives-ouvertes.fr/hal-01666379
DOI : 10.1016/j.epsl.2017.10.009

Éditeur(s) : HAL CCSD Elsevier
Résumé : International audience
During the last decade, the interaction of deep processes in the lithosphere and mantle with surface processes (erosion, climate, sea-level, subsidence, glacio-isostatic readjustment) has been the subject of heated discussion. The use of a multidisciplinary approach linking geology, geophysics, geodesy, modelling, and geotechnology has led to the awareness of coupled deep and surface processes. Deep earth dynamics (topography, erosion, tectonics) are strongly connected to natural hazards such as earthquakes, landslides, and tsunamis; sedimentary mass transfers have important consequences on isostatic movements and on georesources, geothermal energy repartitions. The ability to read and understand the link between deep Earth dynamics and surface processes has therefore important societal impacts. Ground-truthing at carefully-selected sites of investigation is imperative to better understand these connections.Due to its youth (<30 Ma) and its subsidence history, the almost land-locked Gulf of Lion–Sardinia continental margins system provides a unique record of sedimentary deposition from the Miocene to present. Due to its high subsidence rate, palaeoclimatic variations, tectonic events and vertical evolution are all recorded here at very high resolution. The late Miocene isolation and desiccation of the Mediterranean, the youngest and most catastrophic event, the Messinian Salinity Crisis (MSC), induced drastic changes in marine environments: widespread deposition of evaporite (gypsum, anhydrite and halite) in the central basin, and intense subaerial erosion along its periphery. These extraordinary mass transfers from land to sea induced strong isostatic re-adjustments that are archived in the sedimentary record and represent a window to the lithospheric rheology and the deep processes.The GOLD (Gulf of Lion Drilling) project, proposes to explore this unique sedimentary record as well as the nature of the deep crustal structure, providing valuable information about the mechanisms underlying vertical motions in basins and their margins.
ISSN: 0264-8172

insu-01197393
https://hal-insu.archives-ouvertes.fr/insu-01197393
https://hal-insu.archives-ouvertes.fr/insu-01197393/document
https://hal-insu.archives-ouvertes.fr/insu-01197393/file/Rabineau%20et%20al.-final-MPG2014_REVISED-FINAL-No-Marked_FINAL%2BFigures.pdf
DOI : 10.1016/j.marpetgeo.2015.03.018

Éditeur(s) : HAL CCSD American Geophysical Union
Résumé : International audience
Probabilistic seismic hazard analyses (PSHA) are commonly based on frequency magnitude statistics from 50-100 yearlong earthquake catalogs, assuming that these statistics are representative of the longer-term frequency of large earthquakes. We test an alternative PSHA approach in continental western Canada, including adjacent areas of northwestern U. S. A., using regional strain rates derived from 179 Global Positioning System (GPS) horizontal velocities. GPS strain rates are converted to earthquake statistics, seismic moment rates, and ground shaking probabilities in seismic source zones using a logic-tree method for uncertainty propagation. Median GPS-based moment rates and shaking estimates agree well with those derived from earthquake catalogs in only two zones (Puget Sound and mid-Vancouver Island). In most other zones, median GPS-based moment rates are 6-150 times larger than those derived from earthquake catalogs (shaking estimates 2-5 times larger), although the GPS-based and catalog estimates commonly agree within their 67% uncertainties. This discrepancy may represent an under-sampling of long-term moment rates and shaking by earthquake catalogs in some zones; however a systematic under-sampling is unlikely over our entire study area. Although not demonstrated with a high confidence level, long-term regional aseismic deformation may account for a significant part of the GPS/catalog discrepancy and, in some areas, represent as much as 90% of the total deformation budget. In order to integrate GPS strain rates in PSHA models, seismic versus aseismic partitioning of long-term deformation needs to be quantified and understood in terms of the underlying mechanical processes.
ISSN: 0148-0227

hal-00689004
https://hal.archives-ouvertes.fr/hal-00689004
DOI : 10.1029/2011JB008213

Éditeur(s) : Elsevier
Résumé : We describe for the first time a set of large active thrusts and folds near the foot of the easternmost Algerian margin, Western Mediterranean, from swath bathymetry and high-resolution seismic data acquired in 2005 during the Maradja2/Samra cruise. This active system resumes a previous passive margin and creates growth strata deposition on the limbs of large folds, resulting in the development of perched basins at the foot of the margin since less than similar to 1 Ma. They form a set of overlapping fault segments verging toward the Algerian basin, in a way similar to what has been observed off eastern Algiers on the rupture zone of the 2003 Mw 6.8 Boumerdes earthquake. The horizontal shortening rate across large folds is estimated to be of the order of 1 mm/yr. Although no historical earthquakes are reported here, these fault segments could have been responsible for large (M similar to 7.5) events in the past. This young tectonic system further supports the hypothesis of subduction inception of the Neogene oceanic lithosphere in the context of the Africa-Eurasia convergence.
Marine Geology (0025-3227) (Elsevier), 2009-06 , Vol. 261 , N. 1-4 , P. 17-32

Droits : 2009 Published by Elsevier Inc.
http://archimer.ifremer.fr/doc/2009/publication-6850.pdf
DOI:10.1016/j.margeo.2008.05.016
http://archimer.ifremer.fr/doc/00000/6850/

Éditeur(s) : HAL CCSD Elsevier
Résumé : International audience
Lithospheric-scale faults related to oblique subduction are responsible for some of the most hazardous earthquakes reported worldwide. The mega-thrust in the Sunda sector of the Sumatran oblique subduction has been intensively studied, especially after the infamous 2004 Mw 9.1 earthquake, but its onshore kinematic complement within the Sumatran subduction, the transform Sumatran Fault System, has received considerably less attention. In this paper, we apply a combination of analysis of Digital Elevation Models (ASTER GDEM) and field evidence to resolve the kinematics of the leading edge of deformation of the northern sector of the Sumatran Fault System. To this end, we mapped the northernmost tip of Sumatra, including the islands to the northwest, between 4.5°N and 6°N. Here, major topographic highs are related to different faults. Using field evidence and our GDEM structural mapping, we can show that in the area where the fault bifurcates into two fault strands, two independent kinematic regimes evolve, both consistent with the large-scale framework of the Sumatran Fault System. Whereas the eastern branch is a classic Riedel system, the western branch features a fold-and-thrust belt. The latter contractional feature accommodated significant amounts (c. 20%) of shortening of the system in the study area. Our field observations of the tip of the NSFS match a strain pattern with a western contractional domain (Pulau Weh thrust splay) and an eastern extensional domain (Pulau Aceh Riedel system), which are together characteristic of the tip of a propagating strike-slip fault, from a mechanical viewpoint. For the first time, we describe the strain partitioning resulting from the propagation of the NSFS in Sumatra mainland. Our study helps understanding complex kinematics of an evolving strike-slip system, and stresses the importance of field studies in addition to remote sensing and geophysical studies.
ISSN: 0040-1951

hal-01467196
https://hal.archives-ouvertes.fr/hal-01467196
DOI : 10.1016/j.tecto.2016.04.050

Éditeur(s) : HAL CCSD American Geophysical Union
Résumé : International audience
The majority of the largest subduction megathrust earthquakes share the common characteristic of rupturing more than one asperity along strike of the margin. Understanding the factors that control coseismic failure of multiple asperities, and thus maximum magnitude, is central for seismic hazard assessment. To investigate the role of asperities size and spacing on maximum magnitude, seismicity rate, and percentage of synchronized ruptures, we use analog models simulating along-strike rupture behavior of megathrust earthquakes. We found negative correlations between the barrier-to-asperity length ratio Db/Da and maximum magnitude and seismicity rate. Db/Da also controls the process of asperities synchronization along the megathrust. A permanent barrier behavior is observed for Db/Da > 0.5. Comparing our experimental results to the Nankai Trough historical seismicity, we propose that the distribution of megathrust frictional heterogeneities likely explains the diversity of earthquakes which occurred there.
ISSN: 0094-8276

hal-01685546
https://hal.archives-ouvertes.fr/hal-01685546
https://hal.archives-ouvertes.fr/hal-01685546/document
https://hal.archives-ouvertes.fr/hal-01685546/file/Corbi_et_al-2017-Geophysical_Research_Letters.pdf
DOI : 10.1002/2017GL074182

Éditeur(s) : HAL CCSD Springer Verlag
Résumé : International audience
This paper presents a synthesis of a multidisciplinary study carried out along the Nîmes Roman aqueduct, located in the southeast part of France. The study was motivated by archaeologists attempting to explain the partial destructions of only one aerial bridge of the aqueduct (Pont de la Lône). Given its close proximity to the Nîmes seismically active fault, a possible seismic origin for the destructions was invoked. Seismologists and structural engineers thus carried out a variety of field and numerical investigations to test the seismic hypothesis. Supporting field evidence was found first along the aerial bridge section of the aqueduct: broken stalactites, arch warping, cracks and destruction of the bridge just above a breast wall shortly after its construction. Secondly, the underground part of the canal was analysed: irregularity in the thickness of calcite deposits of the canal walls, presence of numerous cracks, horizontal shift of the otherwise linear structure of the canal and presence of calcite twins in the deposits, found only where the canal crosses the Nîmes fault system. Numerical modelling and experimental results show that (1) only the Pont de la Lône would have suffered serious damage under seismic solicitation (assuming an M6 earthquake at 10 km distance). The second aerial bridge, the three levels arches Pont du Gard, an historical monument still standing today, would have suffered less damage, due to its very different fundamental frequency of around 0.4 Hz, far from the amplified seismic near-field spectrum; (2) the numerical models also show that a fall of the canal wall would require a higher magnitude event (M > 6); (3) the presence of calcite twinning requires a differential static stress of 4 MPa or greater, which can only be achieved very close to the fault rupture of a M > 6 earthquake; (4) finally, local surface rupturing of such a fault would also corroborate the hypothesis that the observed offset of the canal may be partly seismically induced. Although other possible origins for each individual evidence may not be excluded, the observed spatiotemporal concentration of architectural/geological anomalies together with the numerical results allow us to support a possible co-seismic origin for these disorders, indirectly attesting to the potential seismic activity (M > 6) of the nearby Nîmes fault. Furthermore, following the conclusions of archaeological studies, the disorders occurred between 250 and 350 year AD, thus providing timing for this possible seismic event, an essential parameter in seismic hazard analysis. These results validate the archaeo-seismological approach as a tool that may help improve the knowledge of major infrequent earthquakes in areas of moderate seismic activity.
ISSN: 0921-030X

hal-00379064
https://hal.archives-ouvertes.fr/hal-00379064
DOI : 10.1007/s11069-008-9276-9

Éditeur(s) : HAL CCSD Elsevier
Résumé : International audience
The Zandjan province is situated in the NW of Iran, at the junction between the southern termination of the Tabriz right-lateral strike-slip fault and the western end of the Alborz active mountain range. Despite this key location, in terms of active tectonics, there is no destructive earthquake reported in available historical seismicity catalogues. Even if a few low magnitude instrumental earthquakes have been recorded indicating that this area is actively deforming, the Zandjan region can be considered as a seismic gap. Based on detailed field investigations, digital elevation models (DEMs), satellite image and aerial photo interpretations, we have identified a major active fault network affecting this regions. It is mainly constituted by several west and est verging thrust faults, trending NW-SE to N-S, that clearly cut Late Quaternary to Holocene deposits. Evidences in the morphology indicate also that they accommodate some right-lateral slip component. This fault network can be divided into four major fault zones affecting Zandjan city (present day population over 500,000) and part of the Zandjan-Mianeh basin more to the west. Based on fault geometrical characteristics (fault length, segmentation ... ) and kinematics, the Zandjan fault system is certainly capable to produce moderate to large earthquakes (Mw = 6.5-7.0) and serious damages in the Zandjan city and surrounding areas. The study of these faults is, then, of major interest for seismic hazard assessment in Zandjan province where total population exceed 1.5 million but also to better understand present day geodynamic of NW Iran.
ISSN: 0040-1951

hal-00617526
https://hal.archives-ouvertes.fr/hal-00617526
DOI : 10.1016/j.tecto.2011.04.012

Éditeur(s) : HAL CCSD Elsevier
Résumé : International audience
We use velocities from 65 continuous stations and 31 survey-mode GPS sites as well as kinematic modeling to investigate present day deformation along the Africa-Iberia plate boundary zone in the western Mediterranean region. The GPS velocity field shows southwestward motion of the central part of the Rif Mountains in northern Morocco with respect to Africa varying between 3.5 and 4.0 mm/yr, consistent with prior published results. Stations in the southwestern part of the Betic Mountains of southern Spain move west-southwest with respect to Eurasia (similar to 2-3 mm/yr). The western component of Betics motion is consistent with partial transfer of Nubia-Eurasia plate motion into the southern Betics. The southward component of Betics motion with respect to Iberia is kinematically consistent with south to southwest motion of the Rif Mountains with respect to Africa. We use block modeling, constrained by mapped surface faults and seismicity to estimate the geometry and rates of strain accumulation on plate boundary structures. Our preferred plate boundary geometry includes one block between Iberia and Africa including the SW Betics, Alboran Sea, and central Rif. This geometry provides a good fit to the observed motions, suggesting a wide transpressive boundary in the westernmost Mediterranean, with deformation mainly accommodated by the Gloria-Azores fault system to the West and the Rif-Tell lineament to the East. Block boundaries encompass aspects of earlier interpretations suggesting three main deformation styles: (i) extension along the NE-SW trending Trans-Alboran shear zone, (ii) dextral strike-slip in the Betics corresponding to a well defined E-W seismic lineament, and (iii) right lateral strike-slip motion extending West to the Azores and right-lateral motion with compression extending East along the Algerian Tell. We interpret differential motion in the Rif-Alboran-Betic system to be driven both by surface processes related the Africa-Eurasia oblique convergence and sub-crustal dynamic processes associated with the long history of subduction of the Neotethys ocean lithosphere. The dextral slip identified in the Betic Mountains in Southern Spain may be related to the offshore fault that produced the Great 1755 Lisbon Earthquake, and as such may represent a significant seismic hazard for the West Mediterranean region.
ISSN: 0012-821X

hal-00618337
https://hal.archives-ouvertes.fr/hal-00618337
DOI : 10.1016/j.epsl.2011.05.048

Éditeur(s) : HAL CCSD
Résumé : 38 pages
Rapport
The Chicxulub impact crater, México, is unique. It is the only known terrestrial impact structure that has been directly linked to a mass extinction event and the only terrestrial impact with a global ejecta layer. Of the three largest impact structures on Earth, Chicxulub is the best preserved. Chicxulub is also the only known terrestrial impact structure with an intact, unequivocal topographic peak ring. Chicxulub’s role in the Cretaceous/Paleogene (K-Pg) mass extinction and its exceptional state of preservation make it an important natural laboratory for the study of both large impact crater formation on Earth and other planets and the effects of large impacts on the Earth’s environment and ecology. Our understanding of the impact process is far from complete, and despite more than 30 years of intense debate, we are still striving to answer the question as to why this impact was so catastrophic.During International Ocean Discovery Program (IODP) Expedition 364, Paleogene sediments and lithologies that make up the Chicxulub peak ring were cored to investigate (1) the nature and formational mechanism of peak rings, (2) how rocks are weakened during large impacts, (3) the nature and extent of post-impact hydrothermal circulation, (4) the deep biosphere and habitability of the peak ring, and (5) the recovery of life in a sterile zone. Other key targets included sampling the transition through a rare midlatitude section that might include Eocene and Paleocene hyperthermals and/or the Paleocene/Eocene Thermal Maximum (PETM); the composition and character of the impact breccias, melt rocks, and peak-ring rocks; the sedimentology and stratigraphy of the Paleocene–Eocene Chicxulub impact basin infill; the chronology of the peak-ring rocks; and any observations from the core that may help us constrain the volume of dust and climatically active gases released into the stratosphere by this impact. Petrophysical property measurements on the core and wireline logs acquired during Expedition 364 will be used to calibrate geophysical models, including seismic reflection and potential field data, and the integration of all the data will calibrate impact crater models for crater formation and environmental effects. The proposed drilling directly contributes to IODP Science Plan goals:Climate and Ocean Change: How resilient is the ocean to chemical perturbations? The Chicxulub impact represents an external forcing event that caused a 75% level mass extinction. The impact basin may also record key hyperthermals within the Paleogene.Biosphere Frontiers: What are the origin, composition, and global significance of subseafloor communities? What are the limits of life in the subseafloor? How sensitive are ecosystems and biodiversity to environmental change? Impact craters can create habitats for subsurface life, and Chicxulub may provide information on potential habitats for life, including extremophiles, on the early Earth and other planetary bodies. Paleontological and geochemical studies at ground zero will document how large impacts affect ecosystems and effects on biodiversity.Earth Connections/Earth in Motion: What are the composition, structure and dynamics of Earth’s upper mantle? What mechanisms control the occurrence of destructive earthquakes, landslides, and tsunami? Mantle uplift in response to impacts provides insight into dynamics that differ between Earth and other rocky planets. Impacts generate earthquakes, landslides, and tsunami, and scales that generally exceed plate tectonic processes yield insight into effects, the geologic record, and potential hazards.IODP Expedition 364 was a Mission Specific Platform expedition to obtain subseabed samples and downhole logging measurements from the sedimentary cover sequence and peak ring of the Chicxulub impact crater. A single borehole was drilled into the Chicxulub impact crater on the Yucatán continental shelf, recovering core from 505.7 to 1334.73 m below seafloor with ~99% core recovery and acquiring downhole logs for the entire depth.
International Ocean Discovery Program

hal-01639520
https://hal.archives-ouvertes.fr/hal-01639520
DOI : 10.14379/iodp.pr.364.2017

Éditeur(s) : HAL CCSD American Geophysical Union
Résumé : Giant earthquakes (Mw≥8.5) usually occur on the boundary between subducting and overriding plates of converging margins, but it is not yet clear which (if any) subduction zones are more prone to produce such a kind of events. Here we show that subduction zones may have different capabilities to produce giant earthquakes. We analyze the frequency-magnitude distribution of the interplate earthquakes at subduction zones that occurred during 1976–2007 and calculate the propensity (defined as the average annual rate) of giant events for about half of the subduction zones. We find that the b value of interplate earthquakes is significantly different among the subduction zones, and out-of-sample giant earthquakes (before 1976 and after 2007) have occurred preferentially in high-propensity areas. Besides the importance for seismic hazard assessment and risk mitigation, our results seem to indicate that a higher seismicity rate does not necessarily imply a higher likelihood to generate giant earthquakes, and the way in which the stress is released at a subduction interface does not change significantly after the occurrence of such events.
ISSN: 0148-0227

hal-01467738
https://hal.archives-ouvertes.fr/hal-01467738
DOI : 10.1002/2016JB013054