Original Paper Landslides DOI 10.1007/s10346-015-0674-2 Received: 26 November 2014 Accepted: 20 December 2015 © Springer-Verlag Berlin Heidelberg 2016

D. Leynaud I T. Mulder I V. Hanquiez I E. Gonthier I A. Régert

Sediment failure types, preconditions and triggering factors in the Gulf of Cadiz

Abstract A series of morphological structures, such as scars and escarpments related to seafloor instabilities, were observed in the Gulf of Cadiz using multibeam bathymetry and acoustic imagery. According to the geometry of the slide scars, the slope angle, the surrounding seafloor morphology and the mechanical parameters of the sediment, we suggest the likely mechanisms initiating the failures for the different types of observed structures. Most of the small-scale sediment failures (≤2 km2) seem directly related to dome-like structures (where slopes are steep) or are located in the vicinity of such structures (fluid flows). It appears that progressive deformation or fluid flow related to the growing of domelike structures may have weakened the sediments sufficiently to bring 7°-steep slopes to metastable conditions (with a factor of safety close to 1.0). The other instability types are likely related to high-magnitude (Ms > 6) earthquakes, which are prone to occur in this area (located in the neighbourhood of the 1755 Lisbon earthquake area). Some particular large-scale structures were observed among these seafloor features, for example on the Guadalquivir Bank. On this bank, a series of successive large scars (at least 4 km long), composed of multiple and very regular arcuate segments (1 km in diameter), were observed at different bathymetric levels (every 40 m). These structures might be related to a deep-rooted detachment zone (e.g. successive listric faults) and triggered by high-magnitude earthquakes or by accumulated displacement along a tectonic discontinuity. This would explain such a largescale deformation, providing a regular escarpment of 40 m high without any sediment flow downslope, thereby suggesting an ongoing (or unfinished) deformation. Keywords Sediment failures . Gulf of Cadiz . Stability analysis . Mud volcanoes . Infinite slope analysis Introduction Seafloor instabilities are commonly observed where slope gradients increase (i.e. shelf break) and in areas of rapid sedimentation (river mouth, delta front, Mulder et al. 2009; glaciated margin, Kvalstad et al. 2005). The most commonly suggested triggering mechanisms at the origin of instabilities are sediment overloading and slope oversteepening. In shallow marine environments, such as the continental shelf, many oceanic processes can also contribute to trigger instabilities, through an increase of pore pressure (swell and wave stress), by increasing sediment load and/or under the action of oceanic processes such as a strong coastal drift (Piper and Normark 2009). Shaking from earthquakes can also play an important role in triggering instabilities; for example, the historical magnitude 7.2 Grand Banks earthquake generated multiple shallow scars along the upper St. Pierre slope in the neighbourhood of the Laurentian Channel and initiated a large turbidity current that flowed to the Sohm Abyssal Plain (Heezen and Ewing 1952). Seafloor failures are instigated in two main ways (Duncan 1996): the downslope increase of shear stress (sediment load, steep

slopes, earthquake shaking) or the decrease of shear strength within the sediments (excess pore pressure or shear strength degradation resulting from cyclic loading due to earthquake shaking, Sultan et al. 2004b). More recently, fluid escape (e.g. water, methane, Sultan et al. 2001; Sultan et al. 2004a) within the seabed was identified as an important factor controlling instabilities (Sultan et al. 2004b). Fluid, including liquid or gas that accumulates in the sediment, can escape under overburden pressure or because of a sudden increase of interstitial pressure (for example, during earthquake shaking). Such overpressure can occur in any submarine environment, including lakes (Chapron et al. 2004) and oceans. Gas escape generates pockmarks in the seafloor and reworks sediments, leading to a zone of weakness. This zone of weakness reduces the shear strength, and increased overpressure may result in sediment failure (McIver 1982; Sultan et al. 2003; Mienert et al. 2004). Pockmarks are usually observed along scarps of large submarine slides, such as the Storegga Slide (Mienert et al. 2005; Hovland et al. 2005). In the Storegga case study, pockmarks are related to gas-hydrate dissociation (Sultan et al. 2003). In the Gulf of Cadiz, fluid escape structures, including pockmarks and collapse structures, have been described recently (León et al. 2006). In particular, these authors described blind valleys located above diapiric structures and generated either by recent distension due to diapirism (León et al. 2010). Due to this broad variety of destabilising factors, a slope failure on the seafloor is a common process that occurs even on gentle slopes. Sediment mass wasting processes are generally subdivided into mass failures (consolidated sliding material) and mass flows (unconsolidated material), which behaves as a fluid (Mulder and Cochonat 1996). Translational mass failures (slides) and rotational mass failures (slump) can be distinguished by the 3D geometry of the bathymetric scar that results from the mass wasting (Leynaud and Sultan 2010). Slides correspond to displacement along a failure plane and generally have a depth-to-length ratio (D/L; D, maximum depth of the slip surface; L, total length of the slip surface; including the part covered by the remnant slide/slump material) lower than 0.1 (Skempton and Hutchinson 1969). Slumps are characterised by a curved-shape (concave upward) failure surface with a D/L ratio between 0.15 and 0.33 (Skempton and Hutchinson 1969). For slumps, the curved failure surface develops in multiple layers of the sediment structure whereas, for the slide, there is generally a single weak layer that defines the failure plane. Slides and slumps (translational and rotational debris slide, respectively, according to the Varnes classification; Hungr et al. 2014) are commonly the result of a retrogressive failure mechanism (a series of successive failures propagating upslope; Mulder et al. 2009). A third structure called a two-wedge slide (Nadim et al. 2003) can also be observed on the seafloor. Such a feature is characterised by its process of failure, and distinguishing it from other types of failures depends on the remnant mass and the specific failure geometry. Failure geometry in a two-wedge slide is a combination of translational slide and collapse with a D/L ratio usually higher than 0.3; the Landslides

Original Paper headwall height is in the order of the failure plane length, and the resulting geometry forms a deep-rooted slide type. This mass wasting type corresponds to a complex slide (two-wedge slide) according to the Varnes classification (Hungr et al. 2014). Our analysis of seafloor morphology distinguishes these three types of failures. The study is applied to the Gulf of Cadiz where a large variety of mass movement is suspected to occur (Mulder et al. 2009) and a good bathymetric dataset exists. The application of a slope stability back-analysis model to this dataset allows us to propose specific triggering factors and preconditions that depend on geometry, location and environment.

Regional setting The Gulf of Cadiz is located in the Atlantic Ocean offshore of SW Iberia and NW Morocco (Fig. 1). The European-African plate convergence in the Gulf of Cadiz results in dextral strike slip along the Azores-Gibraltar plate boundary. Westward migration of the Betic-Rifean Arc emplaced two huge allochthonous masses, the olistostrome unit and the allochthonous unit of the Gulf of Cadiz (AUGC), both on the Guadalquivir Basin. Major faults (strike slip, extensional and thrust), diapirs and active mud volcanoes observed on the seafloor are likely related to this accretionary prism and active subduction zone (Medialdea et al. 2009). As a consequence, this area is a potential source for high-magnitude tsunamigenic earthquakes such as the 1755 Lisbon earthquake

(Zitellini et al. 1999; Fig. 2). According to Silva et al. (2010), the three large, instrumental earthquakes in this area are (1) the 28 February 1969 (Mw = 8), (2) the 12 February 2007 and (3) the 17 December 2009 (Mw = 5.5). The frequency of events (in 1000 years) for a Mw magnitude equal to or exceeding 7.0, 8.0 and 8.75 is, respectively, 9.28, 2.39 and 0.82 in the Gulf of Cadiz (Matias et al. 2013). Materials and methods Bathymetrical and acoustical imagery data Bathymetric and acoustic imaging (reflectivity) data obtained during Cadisar 1 (2001) and Cadisar 2 (2004) cruises on the RV Suroît were used for a morphological analysis of the structures observed on the seabed (Figs. 1 and 3). Bathymetric and acoustic imagery data were acquired with a Simrad EM300 multibeam echosounder (32 kHz) and processed with CARAIBES software (©IFREMER) to clean and grid data. The spatial and vertical bathymetry resolution is 30 m × 30 m and 2 m, respectively, and the imagery spatial resolution is 12.5 m. The acoustic data were corrected for salinity and density effects using 3 CTD (SBE19 probes) and 85 thermoprobes (Sippican). Core data and in situ geotechnical testing Two of the 53 Kullenberg cores collected during these two cruises were used in this study (Fig. 3); both coring sites are located on the

Fig. 1 General bathymetric map of the Gulf of Cadiz showing the main bathymetric structures and the location of the area surveyed during Cadisar 1 and Cadisar 2 cruises (from Zitellini et al. 2009). Tectonic sketch map from Jiménez-Munt et al. (2001)

Landslides

Fig. 2 Tectonic interpretation (Duarte et al. 2010) and earthquake epicentres (for red rectangle area) in the Gulf of Cadiz (NEIC seismic data, magnitude ≥ 3.0, 2000– 2010); AH Albufeira High, CV Cadiz Valley, GB Guadalquivir Bank, SaP Sagres Plateau

north flank of the Albufeira High. Undrained shear strength (Su) measurements were performed at a later stage (a few years later) on both cores (CADI2KS23 and CADI2KS24), using a laboratory shear vane apparatus (Wykeham Farrance Eng. motorised unit) and split cores (vertical surfaces). Unfortunately, according to the sediment unit weight from cores, Su measurements show that the sediment is overconsolidated (overconsolidation ratio (OCR) = 2; usual ratio in marine sediments from the uppermost metres below the seafloor) with a Cσ0u parameter value of 0.35. This is likely due to O

the ageing process of the sediment during storage (5 years) and most likely does not correspond to the real, in situ conditions. Lessard and Mitchell (1985) observed that change in mechanical properties because of ageing can occur even at zero stress. This produces age hardening, i.e. an increase of the remoulded strength and a decrease in the liquidity index probably due to chemical changes. This could compare to early diagenesis in soft sediments. Thus, as the undrained shear strength (Cu) profile versus depth below seafloor (bsf) for the core CADI2KS24 seems not to represent the in situ soil conditions, we use an empirical relationship to approximate the real in situ value of this parameter to perform

basic slope stability evaluations; Skempton and Bjerrum (1957) proposed the following empirical equation, for normally consolidated clays, that relates the ratio σC0u to the plasticity index (PI) O

Cu 0 ¼ 0:11 þ 0:0037  PI σO

ð1Þ

where Cu is the undrained shear strength (kPa), σO′ is the vertical effective stress (kPa) and PI is the plasticity index (%). This means that for normally consolidated clays, the undrained shear strength (Cu) is directly related to the effective overburden pressure (σO′) and PI. Thus, knowing the unit weight profile of the sediments versus depth and the plasticity index, we are able to approximate the most likely undrained shear strength (Cu) profile versus depth for the area considered in this paper (regional profile), assuming a homogeneous distribution of sediments with core depth and normally consolidated clays. Values of PI were obtained using measurements of the Atterberg limits which are less affected by the ageing of the cores Landslides

Original Paper

Fig. 3 Study area showing the different areas where major instabilities were observed and location of Figs. 5, 9, 10, 11, 12 and 13. Studied cores CADI2KS23 and CADI2KS24 are in the frame of Fig. 5. See Fig. 1 for location

(PI = LL − PL, where LL is the liquid limit and PL is the plastic limit, Kinnison 1915; Casagrande 1932) on CADI2KS24 core samples (6.5 m bsf; Casagrande test, Casagrande 1932). Atterberg limits can also change with ageing and drying, but to a lesser extent than shear strength changes. The values of the Atterberg limits used in this study are shown in Table 1. According to the equation proposed by Skempton and Bjerrum (1957) and the plasticity index, the ratio σC0u is found to be equal to O

0.25. This means that the Cu profile with depth can be estimated using the following equation: 0

Cu ¼ 0:25  σO

ð2Þ

Slope stability evaluation Slope stability analyses were performed using two different methods to evaluate potential failure scenarios and to propose the most likely processes for triggering the failure. The infinite slope analysis model was first used for a simplified 2D approach; a more complex code for 3D slope stability analysis, SAMU_3D (Sultan et al. 2007; see supplemental data and Figs. A1 and A2), was used for a more detailed slope stability analysis. A third model, the two-wedge model (Nadim et al. 2003), was introduced here for a better representation of the rupture process for specific slide scars, which do not correspond to classical slide or slump models. This model proposes a failure Table 1 Atterberg limits measured for the CADI2KS24 core sample

Atterberg limits

LL

PL

PI

CADI2KS24 (6.5 m)

67.5

28.5

39

Landslides

plane geometry associated with a collapse mechanism, fitting well with observations made in some places. The output factor of safety (FoS), the ratio of the average resisting shear strength of the sediment along a potential failure surface and the average shear stress along this failure surface were calculated according to these different methods. Infinite slope stability analysis The simplest analysis corresponds to a slab of sediment sliding on a plane parallel to the seafloor. This model is appropriate for translational slides occurring along a glide plane with a low inclination (parallel to the seabed) but is inappropriate for modelling rotational failures (such as slumps). Inasmuch as the slope is considered infinite, interslice forces balance each other and the evaluation of the FoS (defined as the shear strength divided by the mobilised shear stress) becomes much more simple and is defined with the following equation (Nash 1987): FoS ¼

Su γ 0 z  sinβcosβ

ð3Þ

where Su is the undrained shear strength, γ ′ is the submerged unit weight of the sediment, z is the depth of the failure plane and β is the angle of the glide plane. In this paper, to simplify, the Su value (seafloor site scale) and the Cu value (sample scale) will be considered equal to each other. This model provides the first step towards a simple way of evaluating present-day slope stability or the back-analysis of a failed slope. This equation does not apply if there are anomalous fluid pressures or if there is earthquake loading, but when FoS is close to 1, a moderate-magnitude earthquake in the vicinity of the site with such favourable preconditioning is prone to trigger failure.

SAMU_3D The SAMU_3D software (Sultan et al. 2007) allows an accurate 3D slope stability analysis for complex and heterogeneous slopes. This

approach, based on the upper-bound theorem of plasticity (Chen et al. 2001a, b), avoids the simplifications of the limit equilibrium methods (static and kinematic admissibility; Duncan 1996) and

Fig. 4 Two-wedge model; W1 (first block) and W2 (second block) represent the weight of the respective blocks 1 and 2; S1, S2 and S3 represent the average shear resistance along the corresponding slip planes (Nadim et al. 2003). h = total elevation

Landslides

Original Paper proposes a variety of failure surfaces, with a broad range of complex geometrical shapes appropriate for modelling complex bathymetry. The shapes of the arbitrary failure surfaces were defined using eight parameters (see addendum). A new version of the code was recently developed with a probability algorithm which allows us to consider the parameters of uncertainty in the slope stability evaluation, thus resulting in a critical failure surface associated with a probability of failure instead of a safety factor (SAMU_3D_proba; Leynaud and Sultan 2010). Two-wedge model A two-wedge slope stability model (Nadim et al. 2003; Fig. 4), which is more suitable for representing simple geometry compared to SAMU_3D (complex geometry), is used to calculate the FoS for specific slide scar morphologies associated with dome-like structures. This model is used here to mimic specific slide scars for which the head scarp and the failure plane both have a similar length (high headwall); the corresponding mechanism simulates a sliding block (wedge 2 in Fig. 4) associated with a collapsing one (wedge 1 in Fig. 4). Considering the equilibrium of the two wedges and assuming a similar safety factor on the different slip planes, a closed-form solution for the safety factor is obtained with the following equation: FoS ¼

2009; Figs. 3 and 5). The half-rounded structures affect more than 15 km of the northern Albufeira High slope. In this paper, we focus on the shallower slide scar located at 2200 m water depth, between cores CADI2KS23 and CADI2KS24 (Figs. 5 and 6). The scar headwall, located in the vicinity of Kullenberg cores CADI2KS23 and CADI2KS24, is approximately 110 m high and 2 km long, with an average slope of 25°. The sidewalls are steep, with slopes varying from 22° to 31° (Fig. 5c). According to a D/L ratio (Skempton and Hutchinson 1969) of 0.05, where D = 100 m and L = 2000 m, and to the morphological analysis of the bathymetry data (shape of sidewalls and escarpments), for a series of scars in the vicinity, the scar most likely corresponds to a slide with a planar surface for the glide plane (Figs. 5 and 6). Farther east, similar arcuate scars are observed with elongated remnant (or rafted) blocks downslope (Fig. 5). In between these mass wasting zones, a length of approximately 10 km is observed without any failure on the main part of the slope. This area exhibits a pronounced lineament, suggesting a slight escarpment track in development (Fig. 5). These features might correspond to a progressive deformation zone (creeping; steady deformation under a constant shear stress) where future instabilities like the

max Smax cosðβ−αÞ þ Smax 1 2 sinðβ−αÞ þ S3 W 1 sinβcosðβ−αÞ þ W 2 sinα þ X 1 cosβcosðβ−αÞ þ X 2 cosα ð4Þ

where S1max, S2max and S3max are forces representing maximum shear resistances along the sliding planes (maximum shear resistance before failure and not the mobilised shear resistance); W1 and W2 are the submerged weights of wedges 1 and 2, respectively; and X1 and X2 are the horizontal forces resulting from ground accelerations related to earthquake affecting wedges 1 and 2, respectively. Angles α and β are defined in Fig. 4. The slope stability analysis requires morpho-stratigraphic models and geomechanical parameters of the sediment. In the present study area, a succession of clayey layers (20 m thick), with a different shear strength parameter for each of them, is used to build the model based on core observations. Undrained shear strengths were derived from the Skempton and Bjerrum equations mentioned above (Eqs. 1 and 2). Undrained strength parameters are used in the model given clayey sediment type and the strong earthquake activity in this area, which suggests rapid, undrained loading conditions (Matias et al. 2013; the frequency of events in 1000 years for a Mw magnitude equal to or exceeding 7.0 is 9.28). The epicentral distance to the mass wasting zones is estimated to be approximately 100–150 km for the high-magnitude seismic events (Pro et al. 2013; Horseshoe abyssal plain, SW Cape St. Vincent area) and below 10 km for moderate seismic events. The morphology of the slope is reconstructed based on the lateral unfailed slopes, which correspond to simple slopes (2D or 3D) with different angles according to the observed bathymetry. Results Albufeira High (Portimão Valley) A series of coalescing slump scars is observed along the Portimão Valley, on the north flank of the Albufeira High (Mulder et al. Landslides

Fig. 5 Three-dimensional (a) and 2D (b, c) views of the slide scars in the vicinity of the Albufeira High with location of cores CADI2KS23 and CADI2KS24. See Fig. 3 for location

Fig. 6 Unit weight measurements and trend with depth below seafloor for cores CADI2KS23 and CADI3KS24. See Fig. 5 for location

ones observed eastward and westward could occur. A series of moderate earthquakes can obviously play a role in the development of these features in combination with the creeping process.

A slope stability back-analysis was performed using sediment parameters estimated from core CADI2KS24 (from the area not affected by the slide), the reconstructed slope angle existing before

Fig. 7 FoS calculated using the infinite slope model for different Su values in a probable weak layer; the static equilibrium (failure limit) can be reached with a reduction of the undrained shear strength corresponding to the observed sensitivity (Su/2)

Landslides

Original Paper the slide (5°–6° in average) and the morphology of the slide scar (Figs. 5, 6 and 7). The planar surface observed inside the scar suggests the presence of a weak layer at the initiation of the instability event. This assumption was tested using the infinite slope and the SAMU_3D models (Table 2). For a 6° slope, the safety factor is found to be approximately 2.48 (infinite slope) and more than 2.5 (SAMU_3D), which means that the slope is stable with the considered parameters. In this case, the failure requires a weak layer parallel to the seafloor (5° to 7°), at 110 m depth bsf (Fig. 5c), which could correspond to a remoulded sediment (the remoulded undrained shear strength value is considered equivalent to the maximum undrained shear strength, divided by the sensitivity), likely associated with the creeping process, resulting from the progressive set-up of the Albufeira High; a new evaluation of the FoS was performed using this model (Fig. 7). According to the data obtained from cores CADI2KS23 and CADI2KS24, the sensitivity for the first 8 m of clayey sediment fluctuates from 1.4 to 2.6 (Fig. 7). Sensitivity is also affected by ageing and drying, but in a less extent than shear strength parameter; so, we consider here that sensitivity values around 2.0 are not so far from the real in situ values. Considering a layer at a depth of 110 m, uniformly affected by mechanical remoulding, with the maximum sensitivity value observed here (2.6), the safety factor will be drastically affected by the same order of magnitude (infinite slope model). In these conditions, the safety factor is below 1 (the unity meaning equilibrium at the brink of failure) for a 6° infinite 2D slope angle (Fig. 7) and slightly above 1 for a 7° 3D slope angle (SAMU_3D; Table 2, Fig. 8). This means that slope conditions are prone to be in metastable equilibrium, considering a weak layer at 110 m depth (remoulded sediment horizon, sensitivity = 2.6) and a glide plane configuration more or less parallel to the seafloor (between 6° and 7°). Cadiz Valley A spectacular sediment failure scar is observed at 1500 m water depth on the flank of a dome-like structure in the Cadiz Valley (Figs. 3 and 9). The angle of the slope not affected by the slide is approximately 5°. The reconstructed slope in the slide scar provides an angle of approximately 7° due to the dome set-up (Fig. 9). The headwall is approximately 350 m high with a slope angle close to 20°. The sidewalls are smaller, with a height varying from a few tens of metres to nearly 100 m. The width of the slide (distance between sidewalls) is approximately 4 km. Farther downslope, steepness becomes more gentle leading to a regularly rounded disintegrated flow deposit (smooth morphology), with intact blocks appearing in some places (rough morphology). A slope stability back-analysis was performed using the twowedge model, which is more appropriate for representing the observed geometry on the seafloor, considering the ratio (D/L) between the headwall height (D) and the glide plane length (L). Table 2 Factor of safety calculated for the CADI2KS23 slide (Albufeira High) according to the infinite slope and SAMU_3D models for different slope angles

Slope angle

5°

6°

7°

Infinite slope (Su/2.5)

1.19

0.99

0.85

SAMU_3D (Su/2.5)

1.31

–

1.12

SAMU_3D (Su/2.7)

1.27

–

1.09

Landslides

Fig. 8 Two-dimensional views of the critical 3D failure surface of the CADI2KS23 slide (Albufeira High) compared to bathymetric profile: surface projection (upper), cross section in the direction of the slope (middle) and lateral cross section (bottom); distance and depth in metres; SAMU_3D software (Sultan et al. 2007)

Neither the infinite slope model nor the SAMU_3D model normally used to represent complex rounded geometries is appropriate when considering this ratio. The sediment unit weight was obtained from core CADI2KS24 and the strength profile from Skempton and Bjerrum (1957) empirical equation for normally consolidated clays (Eq. 1); seismic terms in Eq. 4 are not taken into account here considering (a) the high uncertainty in properly modelling these effects and (b) the conditions close to the failure process reached on this site with these parameters; the results are shown in Table 3. The headwall heights corresponding to low safety factor values (h = 350–400 m, FoS = 0.89–1.09; Table 3) are in agreement with those observed on the bathymetric map (Fig. 9; h = 350 m). According to the results of the two-wedge model, a 7° slope in this area (corresponding to the average reconstructed slope before sliding in Fig. 9b) could generate metastable conditions under static conditions if the total headscarp height ≥350–400 m. This

Fig. 9 Slide scar and mass transport deposit in the vicinity of a dome-like structure in Cadiz Valley. Three-dimensional views (a, c) and 2D bathymetric profile of the slide scar (b). See Fig. 3 for location

fits very well with the observed slide scar height in the vicinity of the domes, which is mostly higher than 350 m. The measured volume of the failed sediment is estimated to be 3 km 3 (4 km × 4 km × 0.2 km, using no accurate subsurface information). The real headwall is likely slightly higher than 350 m, considering the probable hidden part due to the presence of mass deposits at the foot of the headwall (Fig. 9). It is worth noting that without seismic data in this area, we do not have information about how much of the deposit is left on the sliding plane, nor any confident measurements of morphometric parameters, so we cannot evaluate if the sliding plane observed in the bathymetry is the result of translational (planar surface) or rotational (non-planar surface) sliding. We are aware of the lack of information in our approach regarding the hidden geometry of the glide plane, but this does not lead to a high uncertainty in the results as the whole geometrical model remains more or less similar. Note that the dimensions of the slide scar are very similar to those of the Lolita diapir slide (Figs. 3 and 12). Sagres area The Sagres Plateau exhibits large-scale slide scars at 1200 m below sea level (Figs. 3 and 10). The main scar extends approximately Table 3 Slope stability back-analysis for the Cadiz Valley dome slide

Cadiz Valley dome slide Su1max, Su2max, Su3max

(Su1max) / 2

H = 200

1.90

1.64

H = 300

1.27

1.19

H = 350

1.09

1.02

H = 400

0.95

0.89

Headwall height/Su model (m)

10 km along the margin and is 6 km wide, with a maximum headscarp height reaching 200 m (Fig. 10b). Secondary scars are also present downslope in some places having headwalls approximately 150 m high. The headwalls are rather steep (30°; Fig. 10b), while the slide planes show gentle slopes (ranging from 1° to 4°). The reconstructed slope of the seafloor before the slide event is assumed to be approximately 5°–7° as was found in other places in the Gulf of Cadiz (Portimão Valley). In the vicinity of this large feature (2700 m water depth; Fig. 10a), a series of linear escarpments are observed, oriented N 60° E (azimuth). These escarpments might be related to a different instability process, suggesting the presence of two types of escarpments. The first type located upslope (1200 m water depth) corresponds to the well-known headwall resulting from a slide, while the second one located downslope (2700 m water depth) is likely related to a deep rupture process affecting the sediment over a few hundreds of metres bsf (impact on the seafloor of a tectonic rupture, such as a dip-slip fault mechanism); this will be discussed in detail later in this paper. In this process, the seafloor of the sediment block moving downward is back-tilted (reverse dipping), as this can be observed in the Guadalquivir Bank (Figs. 3 and 13). It is difficult to distinguish one escarpment type from the other using only 2D bathymetry. Interpretation is made easier using a 3D view coupled to bathymetric profiles. Further east, on the other side of the Lagos Valley, similar structures are observed, suggesting displacements occurring along dip-slip fractures (Figs. 3 and 11; A-B bathymetric profile). The direction of the fractures (escarpments) is approximately N 60° E (azimuth), as previously mentioned for the Sagres Plateau. Lolita salt diapir This large structure (diapiric structure related to salt dynamics) is located in the Cadiz Valley and consists of a rounded dome Landslides

Original Paper volcano can collapse not only by gravitational forces but also due to other processes, such as mud eruption or small tremors associated with such an eruption. However, a gravitational process-induced collapse is tested by means of a slope stability assessment and simulated using the two-wedge numerical model; the resulting safety factors are displayed in Table 4 using strengths from regional strength profile obtained from Eq. 2 (Skempton and Bjerrum 1957) with a unit weight profile from CADI2KS24. Values of FoS become critical (close to 1, which means conditions close to failure) with a geometrical rupture model that generates a headwall elevation (h) close to 300 m. This is in agreement with the observed headwall height of the Lolita salt diapir slide and with those observed along similar two-wedge slides located in the Cadiz Valley (southern flank of the Albufeira High). As shown in Table 3, a slight decrease of the undrained shear resistance in the sediment due to a broad range of processes such as deep fluid transfer destructuring the sediment or affecting the consolidation or deformation due to growing of the dome could explain why a slide occurs at a specific place. It is also very likely that an earthquake or reactivation of the salt diapir could act as a triggering mechanism on metastable flanks.

Fig. 10 a Three-dimensional view of the Sagres Plateau exhibiting large-scale slide scars. See Fig. 3 for location. b Two-dimensional bathymetric profiles through the slide scars observed on the Sagres Plateau; profiles exhibit very low slope angles (1° to 4°) corresponding to the glide plane alternating with steeper angles (22° to 30°; slide scars)

(350 m high and 5 km in diameter) deforming the seafloor at a water depth varying between 1250 m (top) and 1600 m (base; Figs. 3 and 12). A steep scar (30°) approximately 250 m high can be observed on the southern flank of the dome. Considering the thickness of the deposits at the base of the dome (at least 50 m thick) and a two-wedge model for the rupture process (Fig. 4), this scar is most likely higher than 250 m (from morphobathymetry analysis). The use of the two-wedge geometrical model is consistent with the rectangular shape of the sidewalls and the steep slope of the headwall, contrasting with the gentle slope of the apparent slide plane. Three types of structure affect the dome flanks: (1) the two-wedge slide mentioned previously, (2) progressive collapse-induced escarpments from 1250 to 1630 m water depth and (3) lineaments of unknown nature (Fig. 12). It is not always evident which of these processes, either slide or collapse through a deep rupture process (dip-slip fault mechanism), is at the origin of the observed escarpment. A mud Landslides

Guadalquivir Bank A particular large-scale tectonic feature mimicking instability was observed on the southeast border of the Guadalquivir Bank (Figs. 3 and 13). This bank extends for more than 10 km and suggests, at first approach, a series of retrogressive slides from the successive positions of the escarpments. Three particularities were observed: (1) the sharp and regularly rounded scallops (coalescent horseshoe shapes) forming the escarpment mimicking a slide scar (these scallops does not indicate the shape of the underlying faults neither any shallow collapse that could enhances the coalescing arcs), (2) the negative slope angle systematically measured on the seafloor downslope from the escarpment and (3) the absence of deposits downslope of the scar. Using 3D images (Fig. 13a), 2D bathymetric profiles (Fig. 13b) and a multichannel seismic profile (Fig. 14; Terrinha et al. 2009), this structure is interpreted as a large sedimentary mass in a progressive vertical deformation state (continuous small deformation, mainly with a vertical component but also with a possible lateral extent southward) or in an unfinished state of deformation (intermittent but repetitive deformation which stopped but is likely to occur again), corresponding to active or stabilised deformation mimicking a creeping process (or even differential compaction). This feature is most likely generated by syn-sedimentary deformation affecting blocks (forming the flank) moving downslope following a normal faulting process (strong sedimentation draping and smoothing the sharp edges of the successive blocks, Fig. 15). This is confirmed by the regular smooth scars observed on the seafloor, contrasting with sharp and irregular scars observed for the more common slides. This large block corresponds to a succession of deep-rooted dip-slip faults generating nested scars on the seafloor. This is consistent with the geometries observed using a segment of multichannel seismic line ARRIFANO 1183-92-04 (Fig. 14; Terrinha et al. 2009). These authors consider that the Guadalquivir basement high is bounded by a NW-SW trending Mesozoic extensional fault that was reactivated as a reverse fault. According to Medialdea et al. (2009), the Guadalquivir Bank corresponds to

Fig. 11 a Three-dimensional view of the eastern part of the Sagres Plateau; the bathymetric profile AB shows a succession of scars and steps related to a dip-slip fault process. See Fig. 3 for location

the limit of the AUGC (the front of the AUGC is located south of the bank) and the dip-slip fault escarpment mentioned here corresponds to an extensional structure; the Guadalquivir area could have been produced by a combination of gravitational collapse (AUGC) and crustal faults (Guadalquivir Bank). Discussion Evidences of slope instabilities can be detected and localised thanks to the presence of scars and/or deposits on the seafloor. Obviously, post-failure sedimentation is prone to hide these surface signs of instantaneous events. In most cases, only recent instabilities can be observed and analysed in detail on the seafloor using bathymetric data and seismic data are required for analysing these mechanisms in detail. To recover as much information as possible about mass transport processes, parameters such as the 3D geometry of the slide scar, the physical parameters measured in the sediments and the environmental conditions derived from the analysis of seafloor morphology should be observed. This approach helps to discriminate different types of triggering Table 4 Factor of safety calculated for the Lolita salt diapir slide using the twowedge model

Headwall height (m)

Fig. 12 Lolita salt diapir. a Three-dimensional view of the salt diapir northward. b Two-dimensional bathymetric profile through the slide scar (N-S). See Fig. 3 for location

Lolita salt diapir slide FoS with Su1max, Su2max, Su3max

FoS with Su1max / 2

H = 200

1.79

1.64

H = 300

1.19

1.09

H = 400

0.89

0.82

Landslides

Original Paper

Fig. 13 a Northward 3D view of the Guadalquivir Bank showing the very regular arcuate successive escarpments between −500 and −800 m water depth; the main escarpment direction rotates progressively from E-W direction (to the south) to N-S direction (to the east). See Fig. 3 for location; b 2D bathymetric profiles; it is interesting to note the systematic negative slope of the steps (seafloor inclination towards the north); the left side of the profile represents the north, while the right side represents the south

mechanisms and preconditioning factors affecting specific slide areas, and it allows us to propose the more appropriate 3D failure shape to explain the observed geometries on the seafloor. Concerning the sediment parameters (undrained shear strength, sensitivity, unit weight), data are scarce (a few cores) and limited to a small portion of the studied area. However, according to empirical relationships representing sediment parameters, the data used do represent sediments of the study area and the uncertainty related to this reduced geotechnical database should not significantly influence the model analysis. As mentioned by Mulder et al. (2009), the coalescence of slide, or slump, scars can suggest a retrogressive process; the coalescing distribution of slides in this area might also be related to the presence of deep sensitive layers (prone to fail), resulting from fluid transfer dissipating downslope and affecting a large area (from the Albufeira High dome to the valley). An initial scar is prone to trigger another upslope slide according to a retrogressive process, but one of the main factors controlling the failure might also be the presence of a more sensitive layer or sliding plane. Landslides

Different mechanisms are likely to trigger the instabilities in the Gulf of Cadiz (Mulder et al. 2009), such as steep slopes, high sedimentation rates, influence of the Mediterranean outflow water (MOW) as an erosive factor, earthquakes (Fig. 2) and upward movement of deep fluids as well as mobile mud and salt deformation. It is worth noting that fluid escape structures in the Gulf of Cadiz are mainly mud volcanoes. From the different mechanisms suggested in this paper, it appears that earthquake activity (Fig. 2) combined with steep slopes and creeping process (and deep fluid seepage at least in specific areas as mud volcanoes) are likely to initiate most of the described instabilities observed in this area. Steep slopes and seepage are the preconditioning factors, and seismic activity is the trigger. In some cases, fluid flows could also be associated with high sedimentation rates. Progressive and even ongoing mass movements (dip-slip fault-induced escarpments), such as those observed at the Guadalquivir Bank or south of the Albufeira High dome (a collapse generating a graben-like structure), are related either to seismicity (Fig. 2) or to progressive tectonic-induced displacement accumulating over time

Fig. 14 Multichannel seismic line TASYO 8 across the Guadalquivir Bank and Guadalquivir Channel (from Medialdea et al. 2009, reproduced with permission from the author)

Fig. 15 Sketch showing an interpretation of the features observed on the Guadalquivir Bank; strong sedimentation from Mediterranean outflow water (MOW) draping and smoothing the sharp edges of the successive blocks moving downslope

periods much longer than decades (Table 5, Fig. 15). Duarte et al. (2010) also observed crescent-shaped features in this area. However, we interpret them as the result of a different mechanism, most likely related to differential compaction generating a scar-like feature without any sediment deposit. One of the main concerns in mapping the types of instabilities in the Gulf of Cadiz is the discrimination between slide scars and fault escarpments. Preliminary observations of the Guadalquivir Bank classified the succession of terraces as retrogressive slide scars, but further observations suggest that terraces were in fact the seafloor moving downward along a succession of faults affecting the Guadalquivir flank, moving with the slight back-tilting mechanism of the seafloor (slow vertical displacements leading to a progressive collapse of the flank without any mass deposit). Similar observations were made south of the Sagres Plateau, suggesting that these mechanisms occur on a large spatial scale throughout this area. Another interesting aspect is the presence of contourites in the Gulf of Cadiz (Sagres Plateau). Many authors have mentioned the occurrence of mass wasting in relation to contourite deposits on the North Atlantic European margin (Bryn et al. 2005; Laberg and Camerlenghi 2008; Leynaud et al. 2009). This might be related to the specific mechanical behaviour of these fine-grained sediments (Laberg and Camerlenghi 2008). Finally, the recurrent direction of the dip-slip fault escarpment (approx. azimuth N 60° E) observed at different places in the Gulf of Cadiz likely corresponds to the ENE-WSW tectonic structures in relation to the NW to WNW direction of the compressional regime (Medialdea et al. 2009), reactivated during a second phase of an extensional regime. Landslides

Original Paper Table 5 Summary of the main escarpments observed in the Gulf of Cadiz with the suggested triggering mechanisms and preconditioning factors according to their geometry and their environment

Location

Volume (km3)

Model of failure

Environment

Triggering mechanism

Preconditioning

Cadiz Valley

3

Two-wedge

Vicinity of a dome

Dome growing/ earthquake

Fluid transfer?

Lolita salt diapir

0.4

Two-wedge

Dome

Dome growing/ earthquake

Fluid transfer?

Albufeira High

–

Dip-slip fault escarpment (DSFE)

Dome

Tectonics

–

CADI2KS23

0.3

Translational slide

Dome

Earthquake

Sediment remoulding/ fluids

Sagres Plateau

6

Translational slide

Plateau

Earthquake

Stratigraphy?

East Sagres Plateau

–

DSFE

Slope

Tectonics

–

Guadalquivir Bank

–

DSFE

AUGC limit

Earthquake tectonic

–

Gorringe Bank

75

Two-wedge

Slope

Earthquake Tectonic

Fault/fluids

Conclusions Three mass wasting types are observed in the Gulf of Cadiz: translational slides, two-wedge slides and slumps. Another structure, often wrongly associated with the mass wasting process, is related to subsidence and extensional tectonics (dip-slip fault displacement). Finally, four types of escarpments can be distinguished in this area (Fig. 16). The apparent fresh geometry of these features on the seafloor suggests that these events are relatively recent, especially considering the high sedimentation rate related to the MOW. Sliding due to sediment overloading is probably facilitated by tectonic deformation and/or moderate to high seismic activity in some places. According to the observed features and to numerical simulations, a series of triggering mechanisms and preconditioning factors are proposed for the

Gulf of Cadiz. Each of these mechanisms affects specific areas (Fig. 17b, Table 5): 1. Steep slopes likely associated with fluid transfer in sensitive layers in areas of mud volcanoes and topographic highs (Cadiz Valley, Lolita salt diapir). 2. Fluid circulation inferred in deep sediments in the vicinity of topographically high areas with a moderate slope angle (Albufeira High). 3. Moderate- to high-magnitude earthquakes (Sagres Plateau). 4. Syn-sedimentary accumulated vertical displacement along a series of faults generating successive escarpments (SE flank of Guadalquivir Bank). According to our knowledge, this scalloped escarpment associated with these faults has not been previously described in the literature.

Fig. 16 Sketch representing the four main slide mechanism types found in the Gulf of Cadiz. The last one (dip-slip fault escarpment) is not a real slide (mimicked slide), as there is no mass transport

Landslides

Fig. 17 a Location and direction of dip-slip fault escarpments (DSFE) observed in the Gulf of Cadiz with the AUGC front (Medialdea et al. 2009). b Localisation of the main mass wasting types and escarpments in the Gulf of Cadiz

It is worth noting that the escarpments observed in the Gulf of Cadiz that fit well with a geometrical rupture of the two-wedge failure model are always higher than 300 m, as suggested by numerical modelling (resulting safety factors are lower than 1 for these heights; Tables 3 and 4). This model proposes a consistent assumption explaining the origin of translational slide scar geometries in the Gulf of Cadiz. Finally, a 3D numerical model (Sultan et al. 2007) provides an opportunity to define very accurate 3D slip surface (referring to plan-view shapes), fitting well with the scars observed on the seafloor and thus producing more realistic safety factors (based on geometry) that correspond to the real geometry (assumed from

bathymetry) and to a more realistic triggering mechanism. Thus, this 3D model could allow to better constrain 3D soil parameters to explain the observed slide. Acknowledgments The post doc contract was funded by the PRES Bordeaux. The authors would like to express their gratitude towards the Suroît crew and technicians of the University of Bordeaux (EPOC). Thanks to the ANR Isis project. We are grateful to Nabil Sultan, who allowed the use of his 3D numerical model for estimating 3D factors of safety mentioned in this work and to T. Medialdea for providing the HR TASYO8 seismic profile with some useful advices and suggestions. Landslides

Original Paper References Bryn P, Berg K, Stoker MS, Haflidason H, Solheim A (2005) Contourites and their relevance for mass wasting along the Mid-Norwegian margin. Mar Pet Geol 22(1– 2):85–96 Casagrande A (1932) Research on the Atterberg limits of soils. Public Roads 12(3):121130–136 Chapron E, Van Rensbergen P, De Batist M, Bek C, Henriet J-P (2004) Fluid-escape feature as a precursor of a large sublacustrine sediment slide in Lake Le Bourget, NW Alps, France. Terra Nov. 16(5):305–311 Chen Z, Wang X, Haberfield C, Yin J, Wang Y (2001a) A three-dimensional slope stability analysis method using the upper bound theorem, part I: theory and methods. Int J Rock Mech Min Sci 38:369–378 Chen Z, Wang J, Yin J, Wang Y, Haberfield C (2001b) A three-dimensional slope stability analysis method using the upper bound theorem, part II: numerical approaches, applications and extensions. Int J Rock Mech Min Sci 38:379–397 Duarte JC, Terrinha P, Rosas FM, Valadares V, Pinheiro LM, Matias L, Magalhães V, Roque C (2010) Crescent-shaped morphotectonic features in the Gulf of Cadiz (offshore SW Iberia). Mar Geol 271:236–249 Duncan JM (1996) Soil slope stability. In: Turner E, Schuster E (eds) Landslides: investigation and mitigation (special report 247). National Academy, Washington, D.C., pp 337–371 Heezen BC, Ewing M (1952) Turbidity currents and submarine slumps, and the 1929 Grand Banks earthquake. Am J Sci 250:849–873 Hovland M, Svensen H, Forsberg CF et al (2005) Complex pockmarks with carbonateridges off mid-Norway: products of sediment degassing. Mar Geol 218:191–206 Hungr O, Leroueil S, Picarelli L (2014) The Varnes classification of landslide types, an update. Landslides 11(2):167–194 Jiménez-Munt I, Fernàndez M, Torne M, Bird P (2001) The transition from linear to diffuse plate boundary in the Azores-Gibraltar region: results from a thin-sheet model. Earth Planet Sci Lett 192:175–189 Kinnison CS (1915) Technological papers of the Bureau of Standards No. 46. A Study of the Atterberg plasticity method. U.S. Department of Commerce, Bureau of Standards, U.S. Government Printing Office, Washington, D.C., p 10 Kvalstad TJ, Andresen L, Forsberg CF, Berg K, Bryn P, Wangen M (2005) The Storegga slide: evaluation of triggering sources and slide mechanics. Mar Pet Geol 22(1-2):245– 256. doi:10.1016/j.marpetgeo.2004.10.019 Laberg JS, Camerlenghi A (2008) The significance of contourites for submarine slope stability. In: Rebesco M, Camerlenghi A (eds) Contourites, vol 60, Developments in sedimentology. Elsevier, The Netherlands, pp 537–556 Lessard G, Mitchell JK (1985) The causes and effects of aging in quick clays. Can Geotech J 22(3):335–346 León R, Somoza L, Medialdea T, Maestro A, Díaz-del-Río V, Fernández-Puga MC (2006) Classification of sea-floor features associated with methane seeps along the Gulf of Cádiz continental margin. Deep-Sea Res II 53:1464–1481 León R, Somoza L, Medialdea T, Hernández-Molina FJ, Vázquez JT, Díaz-del-Rio V, González FJ (2010) Pockmarks, collapses and blind valleys in the Gulf of Cádiz. Geo-Mar Lett 30:231–247 Leynaud D, Mienert J, Vanneste M (2009) Submarine mass movements on glaciated and non-glaciated European continental margins: a review of triggering mechanisms and preconditions to failure. Mar Petrol Geol 26:618–632 Leynaud D, Sultan N (2010) 3-D slope stability analysis: a probability approach applied to the Nice slope (SE France). Mar Geol 269:89–106 Matias LM, Cunha T, Annunziato A, Baptista MA, Carrilho F (2013) Tsunamigenic earthquakes in the Gulf of Cadiz: fault model and recurrence. Nat. Hazards Earth Syst. 13, 1-13, 2013, www.nat-hazards-earth-syst-sci.net/13/1/2013. doi:10.5194/nhess-13-1-2013. McIver RD (1982) Role of naturally occurring gas hydrates in sediment transport. Am Assoc Petr Geol Bull 66:789–792 Medialdea T, Somoza L, Pinheiro LM, Fernandez-Puga MC, Vazquez JT, Leon R, Ivanov MK, Magalhaes V, Diaz-del-Rio V, Vegas R (2009) Tectonics and mud volcano development in the Gulf of Cadiz. Mar Geol 261:48–63 Mienert J et al (2005) Ocean warming and gas hydrate stability on the mid-Norwegian mar-gin at the Storegga Slide. Mar Pet Geol 22:233–244 Mienert J, Vanneste M, Guidard S and Iversen S (2004) NFR Strategisk Universitets prosjekt (SUP)—slope stability. Institutt for Geologi, Universitetet I Tromsø Mulder T, Cochonat P (1996) Classification of offshore mass movements. J Sediment Res 66(1):43–57

Landslides

Mulder T, Gonthier E, Lecroart P, Hanquiez V, Marchès E, Voisset M (2009) Sediment failures and flows in the Gulf of Cadiz (Eastern Atlantic). In: V. Gaullier and B. Vendeville. (eds), Mar. and Petrol. Geol. Spec. Issue, 26: 660–672 Nadim F, Krunic D, Jeanjean P (2003) Probabilistic slope stability analyses of the Sigsbee Escarpment. Proceedings, OTC paper 15203, Offshore Technol. Conf. ’03, Houston, Texas, May 2003, pp. 1-8 Nash D (1987) A comparative review of limit equilibrium methods of stability analysis. In: Anderson, M.G., Richards, K.S. (eds.), Slope stability. John Wiley and Sons, pp. 11-75 Piper DJW, Normark WR (2009) Processes that initiate turbidity currents and their influence on turbidites: a marine geology perspective. J Sediment Res 79(5):347–362 Pro C, Buforn E, Bezzeghoud M, Udias A (2013) The earthquakes of 29 July 2003, 12 February 2007, and 17 December 2009 in the region of Cape Saint Vincent (SW Iberia) and their relation with the 1755 Lisbon earthquake. Tectonophysics 583:16–27. doi:10.1016/j.tecto.2012.10.010 Silva S, Romsdorf M, Matias L, Geissler W, Terrinha P, Carrilho F and NEAREST, W.G. (2010) Characterization of the seismicity in the Gulf of Cadiz based on eleven month monitoring by the NEAREST OBS network EGU General Assembly. Geophys Res Abstr, Vol. 12 Skempton AW, Bjerrum L (1957) A contribution to the settlement analysis of foundations on clay. Geotechnique 7:168 Skempton AW, Hutchinson JN (1969) Stability of natural slopes and embankment foundations, state-of-the-art report. In: Proc. 7th Int. Conf. SMFE, Mexico City, vol. 2, pp. 291-335 Sultan N, Cochonat P, Bourillet JF, Cayocca F, Colliat JL (2001) Analysis of submarine slumping in the Gabon continental slope. AAPG 88(6):781–799 Sultan N, Cochonat P, Foucher JP (2003) Effect of gas hydrate dissociation on seafloor slope stability. Submarine mass movements and their consequences. Kluwer Academic, Dordrecht, pp 103–11 Sultan N, Cochonat P, Foucher JP, Mienert J (2004a) Effect of gas hydrates melting on seafloor slope instability. Mar Geol 213(1):379–401 Sultan N, Cochonat P, Canals M, Cattaneo A, Dennielou, B, Haflidason H, Laberg J, Long D, Mienert J, Trincardi F, Urgeles R, Vorren T, Wilson C (2004b) Triggering mechanisms of slope instability processes and sediment failures on continental margins: a geotechnical approach. Marine Geology, 213(1-4), 291-321. Publisher’s official version: doi:10.1016/ j.margeo.2004.10.011, Open Access version: http://archimer.ifremer.fr/doc/00000/700/ Sultan N, Gaudin M, Berné S, Canals M, Urgeles R, Lafuerza S (2007) Analysis of slope failures in submarine canyon heads: an example from the Gulf of Lions. Journal of Geophysical Research (JGR) - Earth surface, 112(112-F1), NIL_1-NIL_29. Publisher’s official version: doi:10.1029/ 2005JF000408, Open Access version: http://archimer.ifremer.fr/doc/00000/2410/ Terrinha P, Matias L, Vicente J, Duarte J, Luis J, Pinheiro L, Lourenço N, Diez S, Rosas F, Magalhaes V, Valadares V, Zitellini N, Roque C, Mendes Victor L, MATESPRO team (2009) Morphotectonic and strain partitioning at the Iberia-Africa plate boundary from multibeam and seismic reflection data. Mar Geol 267:156–174 Zitellini N, Chierici F, Sartori R, Torelli L (1999) The tectonic source of the 1755 Lisbon earthquake and tsunami. Ann Geofis 42:49–55 Zitellini N, Gracia E, Matias L, Terrinha P, Abreu MA, DeAlteriis G, Henriet JP, Danobeitia JJ, Masson DG, Mulder T, Ramella R, Somoza L, Diez S (2009) The quest for the Africa-Eurasia plate boundary west of the Strait of Gibraltar. Earth and Planet Sci Lett 280:13–50

Electronic supplementary material The online version of this article (doi:10.1007/ s10346-015-0674-2) contains supplementary material, which is available to authorized users. D. Leynaud G-tec s.a.s., Le Vaisseau, 120 Boulevard Amiral Mouchez, 76087, Le Havre Cedex, France D. Leynaud : T. Mulder ()) : V. Hanquiez : E. Gonthier : A. Régert UMR 5805 EPOC, Université de Bordeaux, CS 50023, Allée Geoffroy St-Hilaire, 33615, Talence Cedex, France e-mail: [email protected] A. Régert TOTAL EP/EXPLO/IGEO/DER/DM, GEOREX, Tour Coupole, Office 08E50, Courbevoie, La Défense, 92 400, Paris, France