Available online at www.sciencedirect.com

tters 267 (2008) 453–467www.elsevier.com/locate/epsl

Earth and Planetary Science Le

Nonlinear variations of the physical properties along the southern Ecuadorsubduction channel: Results from depth-migrated seismic data

Alcinoe Calahorrano B.a,⁎, Valentí Sallarèsb, Jean-Yves Collota,Françoise Sagea, Cesar R. Raneroc

a Géosciences Azur, Université de Nice Sophia-Antipolis, Institut de Recherche pour le Développement (IRD), Université Pierre et Marie Curie, Centre National dela Recherche Scientifique (CNRS), BP 48, 06235 Villefranche-sur-Mer cedex, France

b Unidad de Tecnología Marina-CMIMA, Consejo Superior de Investigaciones Científicas (CSIC), Passeig Maritím de laBarceloneta 37-49, 08003, Barcelona, Spain

c Institució Catalana de Recerca i Estudis Avançats (ICREA) at Instituto de Ciencias del Mar-CMIMA, Consejo Superior de Investigaciones Científicas (CSIC),Passeig Maritím de la Barceloneta 37-49, 08003, Barcelona, Spain

Received 15 June 2007; received in revised form 25 October 2007; accepted 29 November 2007

Available onlin

Editor: C.P. Jauparte 1 February 2008

Abstract

We use two high-quality pre-stack depth-migrated multichannel seismic profiles acquired to quantify physical properties variations of underthrustsediments along the first ~30 km of subduction off the erosional southern Ecuadorian margin. Seismic data show three zones along the subductionchannel (referred to as Zones I, II and III) characterized by distinct velocity and velocity-derived physical properties, which are in agreement withvalues estimated from experimental results of deformation in granular media. These three zones result from transformational changes of underthrustsediments governed by fundamentally different physical processes that control their mechanical behavior at increasing confining pressures. Based onour observations and its comparison with experimental results, we argue that the transformations undergone by underthrust sediments as they dip intothe subduction zone are the following: within Zone I, progressively increasing velocity (and decreasing velocity-derived porosity) indicatescontinuous sediment compaction, which must be accompanied by effective fluid drainage along the décollement and/or across the accretionarywedge. The underthrust material is here unconsolidated from a mechanical point of view. Laboratory experiments indicate that the dominantprocesses at this range of pressures are grain rolling, particle rotation and frictional slip at grain contacts. Within Zone II, velocity (and porosity)remains constant for ~16 km (SIS-72) and ~12 km (SIS-18). This suggests undrained conditions resulting in growing fluid overpressure at thesubduction channel. Grain deformation is similar to Zone I. Within Zone III, velocity increases and porosity falls rapidly, indicating sedimentcompaction and subsequent release of over-pressured fluids, where grain deformation is likely to be elastic. This might be the dominant process untilthe grains attain their crushing strength, resulting in granular cataclasis and, eventually, in the collapse of the system. We suggest that over-pressuredfluid release may induce hydrofracturation and it is likely to increase inter-plate coupling down from Zone III.© 2007 Elsevier B.V. All rights reserved.

Keywords: subduction channel; velocity inversion; fluid overpressure; grain deformation

⁎ Corresponding author. Instituto de Ciencias del Mar-CMIMA, ConsejoSuperior de Investigaciones Científicas (CSIC), Passeig Maritím de laBarceloneta 37-49, 08003, Barcelona, Spain. Tel.: +34 93 230 9500; fax: +3493 230 95 55.

E-mail address: alcinoe@cmima.csic.es (A. Calahorrano B.).

0012-821X/$ - see front matter © 2007 Elsevier B.V. All rights reserved.doi:10.1016/j.epsl.2007.11.061

1. Introduction

In convergent margins, when subduction takes place, part ofthe oceanic and continental sediments cumulated in the trench iscommonly dragged with the downgoing plate beneath themargin. This downgoing sediment form the so-called subduc-tion channel (SC), a poorly consolidated and fluid-rich layerthat is structurally squeezed between upper and lower plates

454 A. Calahorrano B. et al. / Earth and Planetary Science Letters 267 (2008) 453–467

(Scholl et al., 1977; Shreve and Cloos, 1986). The SC has beenclearly imaged using multichannel seismic (MCS) data alongtens of kilometers beneath several accretionary margins likeNankai (e.g. Moore et al., 2001; Bangs et al., 2004), Barbados(e.g. Westbrook et al., 1982; Moore and Shipley, 1988) andCascadia (e.g. Clowes et al., 1987; Davis and Hyndman, 1989;Hyndman et al., 1990).

Physical and mechanical properties of subducting materialstrongly influence the shape and tectonic deformation of themargin (Davis et al., 1983; Lallemand et al., 1994). Thesubducting material contains pore-filling fluids in variableamounts depending on the nature and thickness of the under-thrust sediment deposit. In high-permeability conditions, porefluids are expelled as a response to the rising pressure resultingfrom the landward-increasing load of the overriding-plate.Sediment framework will mainly support this load and will beprogressively compacted. In the case that a significant part offluids remain trapped within pores, which typically occurs whensubducting sediments are rapidly buried beneath the margin andpermeability conditions are low, the overburden pressure istransferred into increasing fluid over-pressures along the déco-llement and/or within the sediment column (e.g. Bangs et al.,1990). Fluid pressure variations are believed to play a majorrole in controlling deformation processes and fault dynamicsalong subduction zone megathrusts (Moore, 1989; Le Pichonet al., 1993; Moore and Saffer, 2001; Sage et al., 2006). Suchdeformation processes include frontal accretion, wedge thick-ening by out-of-sequence thrusting, subduction erosion, andunderplating (Cloos and Shreve, 1988). Understanding thephysical behaviour of underthrust material is critical because itcontrols mechanical processes such as inter-plate friction,hydrofracturing (Sibson, 1981), and the location of thedécollement (e.g. Le Pichon et al., 1993; McIntosh and Sen,2000). In addition, they also influence mass and fluid budgets(e;g. Saffer and Bekins, 1998), heat transfer (e. g. Hyndmanet al., 1995) and the down-dip physical and chemicaltransformations of subducted material (e.g. Kastner et al.,1991). These transformations that occur as the plate drivesdeeper into the subduction zone are believed to play animportant role on both the location of the seismogenic zone(e.g., Vrolijk, 1990) and the amount of co-seismic slippropagation (e.g., Moore and Saffer, 2001).

Physical properties of the SC material deeply rely on thekinematics, margin stress regime, sediment supply and watercontent, in combination with the age and crustal structures ofboth the downgoing and overriding-plates. At present day, theknowledge of the physical properties of the SC material islimited to direct measurements of sediment porosity/density andseismic velocity obtained during scientific drilling at the leadingedge of sediment prism (b~4 km from the trench and b2 kmbelow sea floor, bsf). This is the case, for example, of theaccretionary prisms of Nankai (Moore et al., 2001) and Bar-bados (Mascle et al., 1988; Moore et al., 1995), as well as thefrontal prism of Costa Rica (Bolton et al., 2000). In contrast,only few indirect estimates of subduction channel porosity andfluid content at greater depths and distances from thedeformation front (N10 km) are available to date. Some of the

few examples are the accretionary complexes of Barbados(Bangs et al., 1990) and Oregon (Cochrane et al., 1994; Yuanet al., 1994), where SC porosity and fluid pressures have beenestimated based on Normal Move Out (NMO) velocity analysisof MCS data. Likewise, von Huene et al. (1998) used PrestackDepth Migration (PSDM) of MSC data to estimate porosity anddewatering at the accretionary margin of Alaskan. Finally,numerical modeling of consolidation and dewatering based onborehole chemical and physical data measured at the margin'stoe has been also made in Barbados (Stauffer and Bekins, 2001)and Nankai (Saffer and Bekins, 1998).

In this paper, we use MCS data acquired across the erosionalmargin of southern Ecuador during the SISTEUR-2000 survey(Collot et al., 2002) to (1) identify the main structures of themargin, (2) obtain the seismic velocity field by means of PSDM,and correspondent depth-images, (3) calculate velocity-derivedporosity at SC, as well as effective and pore pressure to quantifyfluid overpressure variations and dewatering along some~30 km of subduction, and (4) discuss the causes andimplications of estimated physical properties, both across andalong the strike of the margin.

2. Tectonic setting

At the South Ecuador margin, offshore the Gulf of Gua-yaquil, the Nazca plate subducts eastwards beneath South-America at ~55 mm/yr (Trenkamp et al., 2002) (Fig. 1). Inaddition to normal plate convergence, the Ecuador margin, aspart of the so-called North Andean Block, moves north-east-wards in response to subduction obliquity in Colombia andstrain partitioning of north-western South American plate. TheNorth Andean Block motion occurs along a major NE-trendingright-lateral strike–slip fault system (e.g. Winter et al., 1993;Ego et al., 1996) at a rate of ~6±2 mm/yr (Trenkamp et al.,2002). This motion is believed to have favored the opening theGulf of Guayaquil by N–NE extension (Ego et al., 1996; Wittet al., 2006) (Fig. 1).

Recent work based on MCS and swath bathymetry dataacquired during the SISTEUR survey shows that the Ecuadormargin is dominantly erosional and characterized by extension(Collot et al., 2002; Calahorrano, 2005; Sage et al., 2006).Despite this overall erosional behaviour, the margin off the Gulfof Guayaquil is fronted by a ~3–10 km-wide sediment prism.The volume and nature of terrestrial sediment supplied to thetrench or conveyed by the Nazca plate vary considerably alongthis area. This variability is mainly associated to the erosion andsediment transport of the Andes, the presence of the GrijalvaFracture Zone (GFZ) and the Carnegie Ridge, and the marinecurrents. The erosion of the AndeanWestern Cordillera providesseveral hundred meters of terrigenous sediments transportednorthwards to the trench by the Esmeraldas river, and south-wards by the Guayas river up to the Gulf of Guayaquil (Fig. 1).Sand turbidites with abundant wood fragments characterizingthe surficial trench deposits in the northern margin (Collot et al.,2005) corroborate the erosion/transport of clastic material, andillustrate sediment variation along the trench while contrastingthis deposits with dark-green hemipelagic mud with high

Fig. 1. Multibeambathymetry of the Ecuadorianmargin. Black lines are the track of SIS-72 and SIS-18, located off the Gulf of Guayaquil.White lines indicate the portion ofprofile used in this study. Black dash line corresponds to the East limit of the North Andean Block (NAB). Thin dash lines correspond to the two major paths of sedimenttransport to the trench: the Esmeraldas River (ER) andGuayasRiver (GR). Carnegie Ridge is subducting almost perpendicularly below the central-north Ecuadorianmargin,while the Grijalva Fracture Zone (GFZ) subducts obliquely under the Gulf of Guayaquil. NF shows normal, east-dipping faults in the southern flank of the Carnegie Ridge.

455A. Calahorrano B. et al. / Earth and Planetary Science Letters 267 (2008) 453–467

organic content, ash-rich turbidites and basal graded units ofreworked foraminifera in the central margin (Lonsdale, 1978).

The GFZ and the Carnegie Ridge are two major structuralfeatures of the Nazca Plate that enter the Ecuador trench. TheGFZ is a NE-trending, 700 m-high scarp that separates the ~32-34 Ma, smooth morphology of the ancient Farallon lithosphereto the south (Lonsdale and Klitgord, 1978; Barckhausen et al.,2001), from the younger (~0–24 Ma) and rougher Nazcalithosphere to the north (Handschumacher, 1976; Hey, 1977).Sediment accumulation is favored south of the GFZ-trenchintersection where the trench is the deepest (Fig. 1). The Car-negie Ridge, located north of the GFZ, is a prominent ~2 km-high, ~200 km-wide, EW-trending volcanic ridge resultingfrom the interaction between the Galapagos hotspot and the

Nazca–Cocos spreading centre since ~20 Ma (Sallarès et al.,2005; Lonsdale and Klitgord, 1978; Sallarès and Charvis,2003). The southern flank of the Carnegie Ridge is spotted bynumerous seamounts that collide with the Ecuador marginsupplying the trench with abundant mass-wasting deposits(Sage et al., 2006). The cores of site 157 from DSDP Leg 16,and sites 1238 and 1239 from ODP Leg 202 in the southernflank of the easternmost Carnegie Ridge indicate that in thisregion, the basalt basement is draped by a ~400–500 m thicksediment sequence mainly corresponding to diatom nannofossilooze and nannofossil diatom ooze, with varying abundance ofclay, foraminifers siliceous microfossils, and some localconcentration of organic carbon, carbonate and ash layers.Chert and micrite, indicate diagenesis of opal and carbonate at

Fig. 2. Line SIS-72. A) PSDM image of the front of the margin. B) Geological interpretation of seismic line. TSC, Top of subduction channel reflection. TOC, Top ofoceanic crust reflection. C) 2D velocity model. Notice yellowish green colors along the subduction channel, denoting the velocity inversion.

456 A. Calahorrano B. et al. / Earth and Planetary Science Letters 267 (2008) 453–467

the base of the pelagic sequence. Additionally, the CarnegieRidge is a physical barrier controlling the circulation of themarine currents that flow along the trench affecting thesediment deposition and erosion (Lonsdale, 1976).

3. Data

During the SISTEUR survey, some ~6000 km of deepMCS lines were collected using a 45-L airgun source tuned ina single-bubble mode, and a 360-channel, 4.5 km-longstreamer (Collot et al., 2002). Airgun shots were fired every50 m to provide 45-fold common depth point data. Here weuse lines SIS-18 and SIS-72, which extend perpendicularly tothe southern Ecuadorian trench, north and south of the GFZrespectively (Fig. 1). These lines were processed up toPrestack Depth Migration (PSDM) to obtain depth-images(e.g. Fig. 2A) and accurate 2D seismic velocity model (e.g.Fig. 2B).

4. PSDM and velocity model uncertainties

This study focuses on the frontal margin, beneath the lowerslope, which is a complex target for seismic imaging because itusually shows strong lateral and vertical velocity variations. Toovercome difficulties related to the inherent complexity of themedium, we performed PSDM (Yilmaz, 2001), which is atpresent day the most effective method to accurately image thesubsoil considering its spatial heterogeneities (Guo and Fagin,2002). We used the SIRIUS-2.0 software package (GXTechnology) to perform PSDM based on the Kirchhoffalgorithm. This method allows building 2D velocity modelsby focus analysis and iterative velocity picking. The resolutionand accuracy of the obtained velocity model depend basicallyon 1) the quality and dip of reflections, and 2) the maximumsource–receiver offset/depth ratio and the dominant sourcefrequency (Lines, 1993; Ross, 1994). Bartolomé et al (2005)showed that coherent velocity changes for depth-focusing

Fig. 3. Line SIS-18. A) PSDM image of the front of the margin. B) Geological interpretation. C) 2D velocity model. Note that velocity along the subduction channel isslightly lower than that of the overlaying upper-plate basement.

457A. Calahorrano B. et al. / Earth and Planetary Science Letters 267 (2008) 453–467

analysis can be observed until ~15 km depth, when using thesame seismic source and acquisition system that we use here.Consequently, this depth can be considered as a sort of limit forreliable velocity modeling. In lines SIS-72 and SIS-18 thedeepest reflections outlining the subduction channel are strongand continuous, facilitating the application of the PSDMmethod up to ~7 km depth (e.g. Figs. 2A and 3A). To furtherverify the accuracy of the obtained velocity model and constrainthe errors, we performed an uncertainty test using the VelocityScan module of SIRIUS-2.0. This module generates a set ofvelocity profiles by increasingly perturbing the reference modelwithin the region of interest. A total of 16 velocity models, witha ±5% step with respect to the reference velocity model, wereused to generate the corresponding depth-migrated images andcorrected Common Reflection Point (CRP) gathers. CRPgathers enable estimating gather flattening that affects in turnthe quality of depth sections. The comparison of the imagesobtained with the different perturbed velocity models allowsestimating the range of accuracy of the velocity models

considered. The results indicate that, depending on the segmentof the profile, velocity perturbed by more than 5–10% of theoriginal velocity do not flatten CRP gather reflections andreduce considerably the quality of migrated images. Thus, themaximum velocity uncertainties range between ±85 m/s at thetrench axis, where the average SC velocity is ~1700 m/s, to lessthan ±200 m/s at a ~5 km bsf and ~32 km landward from thetrench, where the velocity is ~4000 m/s. This velocityuncertainty is small considering the depth of the target, andsupports the reliability of the velocity inversion at the SC,serving as a base to depth-migrate seismic data and to associateuncertainty bounds to the velocity-related physical properties ofSC material. An ultimate velocity validation was done whenAgudelo (2005) performed a new depth migration of line SIS-72 using the ray+Born pre-strack depth migration/inversionmethod (Thierry et al., 1999). He obtained a similar velocitydistribution, particularly along the SC, and equivalent depthimages that corroborate the accuracy of the velocity model usedin this work.

458 A. Calahorrano B. et al. / Earth and Planetary Science Letters 267 (2008) 453–467

5. Imaging the margin structure and subduction channel

5.1. Line SIS-72

This line images a small frontal prism, the overriding-platebasement and slope sediment, and a well-defined subductionchannel (Fig. 2A and B). Fig. 2A is remarkable because it clearlyshows the SC extending from the trench to ~30 km landward, inan erosional margin.

At the toe of the margin, a ~8-km-wide and ~1 km-thickfrontal prism is formed by three main landward-dipping im-bricated thrust sheets abutting at the apex of the overriding-platebasement. The top of the basement is associated to a strong butdiscontinuous seaward-dipping reflection that underlies thestratified reflections of the slope sediment. The overriding-platebasement is cut by east-dipping reflections probably related toancient faults.

The deformation front is marked by a thin reflection of thecurrent thrust separating the frontal-prism sediments and thehorizontal reflections of trench infill. This reflection continuesbeneath the frontal prism and branches landward to a thick,strong and continuous reflection interpreted as the décollement(Fig. 2A and C). The décollement reflection undulates and dipsto the East marking the top of the subduction channel andbounding the underside of the overriding-plate basement. It iscalled TSC (Top of the subduction channel) reflection from hereon. Seaward dipping reflectors at the bottom part of the upper-plate basement appear to be truncated by the TSC, supportingbasal tectonic erosion. P-wave seismic velocity (Vp) of theupper-plate basement backstop (3.5–3.8 km/s) contrasts with1.8–2.7 km/s of the frontal prism (Fig. 2B and C), which resultsfrom off-scraping sediment of the 0.6 km-thick wedge of well-stratified, flat-laying trench infill deposits. These deposits areinterpreted as turbidites, due to its proximity to the dischargezone of submarine canyons, and by analogy with sand turbiditicdeposits off the northern margin (Collot et al., 2005). Theseturbidites show a Vp=1.8 km/s and lie unconformably over aneast-dipping, 0.2–0.4-km-thick layer with stratify to transparentfacies of pelagic to hemipelagic sediment that drapes thePaleogene oceanic crust (Fig. 2).

The top of the oceanic crust is marked by a clear reflection(TOC, Top of the Oceanic Crust reflection) that extends underthe margin marking the base of the SC (Fig. 2A). The TOCshows 4–5 km-wide, 0.4–0.5 km-relief highs and lowsrepresenting subducting horsts and grabens, suggesting thevertical reactivation of oceanic crust normal faults (e.g. Raneroet al., 2003). Fig. 1 shows some of these east-dipping normalfaults in the southern flank of the Carnegie Ridge.

The SC, bounded at top and bottom by the TSC and theTOC, dips ~4° landward for about 32 km from the trench, anddown to a depth of ~5 km bsf. Its thickness varies from lessthan 0.2 km beneath the frontal prism, to ~0.8–0.9 km on mostof the section before thinning to only ~0.2–0.3 km between29 km to ~32 km from the trench. The origin of thesethickness variations is unconstrained, but we suggest that theymay arise from changes in sediment input from the submarinecanyons located nearby.

The SC shows a well-constrained low velocity (2.7–2.8 km/s)that produces a velocity inversion when compared with the highervelocities (3.5–3.8 km/s) of the overlaying upper-plate basement(Fig. 2C). This velocity inversion results from the contrast betweenthe basement density and the low density of unconsolidated sub-ducting material, and it use to be accompanied by a phase in-version of seismic data polarity (e.g. Bangs et al., 1999). Overpressured fluids are frequently relatedwith these phenomena, but itis not always the case (e.g. Shipley et al., 1990). Nevertheless, inour data, the TSC reflection shows discontinuous polarityinversion and changing amplitude that could be associated withvariations in water content, as suggested by Bangs et al. (2004).

5.2. Line SIS-18

Line SIS-18, located north of the GFZ-trench intersection(Fig. 1), shows different structures and seismic facies whencompared with line SIS-72. Line SIS-18 characterizes anerosional margin, fronted by a ~3-km-wide sediment prism withlittle internal structure suggesting slope debris accumulated atthe deformation front (Fig. 3A and B).

The flat-bottom trench is underlain by a graben formed by anormal fault probably caused by plate bending. There, the trench istwice narrower and shallower than in line SIS-72. Its ~0.3 km-thick infill deposit mainly shows structure-less seismic facies, andan average Vp of 2.0–2.2 km/s (Fig. 3C). The Neogene oceaniccrust is covered by a thin layer of pelagic to hemipelagic sedimentsb0.2 km that is hardly visible at the base of the trench. The chaoticdeposits filling the trench may be related with a semicircular scarobserved in the inner trench slope bathymetry and with an abruptseaward thinning and truncation of the ~0.5 km-thick slopesediment near CDP 9200, to support mass wasting of the lowerslope as the trench infill mechanism (Fig. 3A).

Similarly to line SIS-72, the TOC and TSC reflections delineaterespectively the top of the oceanic crust and the décollement, thusclearly defining the SC. The overriding-plate basement isinternally poorly reflective with a few faint reflections dippingeither landward or seaward, occasionally parallel to the TSC. TheSC dips landward at ~6° and is thinner than in line SIS-72, varyingfrom ~0.2 km beneath the frontal prism to a maximum of ~0.6 kmlandward. Here, changes in the slope sediment thickness indicatethat mass wasting may cause changes in trench sediment input.The SC shows higher velocity than on line SIS-72, and a weak Vpinversion of ~200 m/s with respect to the overriding-platebasement velocity, which is within the Vp uncertainty range(Fig. 3C). The different range of velocity within the SC betweenlines SIS-18 and SIS-72 suggest a different nature of underthrustmaterial, and agrees with the observation of seismic images.

6. Quantifying fluid pressure within the SC

Estimating physical properties of the SC material to quantifyfluid pressure variations requires: 1) to infer the nature of under-thursting sediment based on the geological setting and coringinformation, 2) to use PSDM seismic velocity (Vp) along the SCand sediment composition to infer porosity (Φ) based on existingVp-Φ relationships, 3) to calculate fluid content and fluid flow

Fig. 5. Estimated porosity obtained from velocity considering a Φ0=0.31 andfsh=20, 40, 60 and 80.

Table 1Initial porosity (Φ0), shale fraction (fsh) range, average fsh and b value of linesSIS-72 and SIS-18 used for pressure estimation

Line Φ0 fsh range Average fsh b(MPa−1)

SIS-72 0.45 0.30–0.50 0.40 0.052SIS-18 0.30 0.20–0.30 0.25 0.04

Critical porosity (Φc) is 0.31 as considered in the normal compaction equation ofErickson and Jarrard (1998).

Table 2Acronyms of parameters and calculated physical properties

Vp Seismic P-wave velocityΦ Porosity

459A. Calahorrano B. et al. / Earth and Planetary Science Letters 267 (2008) 453–467

along the different segments of the SC, and 4) to calculate the fluidpressure and its variations within the SC. For these calculations,our first assumption was to consider the SC as a zone decoupledfrom the overriding-plate stress regime by the décollement (LePichon et al., 1993). This decoupling, shown by drilling andMCSdata in other convergent margins (e.g. Westbrook et al., 1982;Taira et al., 1992), implies that below the décollement (i.e. withinthe SC), the horizontal stress resulting from plate convergence islargely exceeded by the vertical stress from lithostatic pressure,and in consequence it can be neglected in the calculations.

We based on a portion of PSDM velocity models of lines SIS-72 and SIS-18 to obtain the average velocity value within the SCfrom the trench to ~30 and ~16 km landward (Figs. 2C and 3C).Velocity and velocity-derived physical properties were calculatedat 100 m interval and subsequently smoothed by applying a 1 km-wide Gaussian-type sliding window to filter high-frequencyartifacts. Then, the first step was to calculate porosity using theVp-Φ empirical relationship for normally compacted water-saturated siliciclastic sediments fromErickson and Jarrard (1998).We use this relationship because it accurately reproduces the Vp-Φ of a large number of sediment samples with lithologies rangingfrom sand to shale acquired in different geodynamic settings (e.g.Erickson and Jarrard, 1998; Gettemy and Tobin, 2003). Thisequation considers sediment compaction without shortening, andtakes into account the critical porosity (Φo) and the shale fraction(fsh) of incoming sediments. Nur et al. (1998) define the Φo as thetransition from the suspension domain to the consolidate rockdomain. The suspension domain, for high-porosity rocks, corres-ponds to a media that is unconsolidated and mechanically fluid-supported. In contrast, the consolidated domain, for low-porosityrocks, corresponds to a media that shows a continuous frame-supportedmatrix. It is important to note that the Vp-Φ relationshipbehaves differently in each domain: when ΦNΦo, velocity and

Fig. 4. Cartoon showing the evolution of porosity (Φ) and thickness (h) of asubducting element from point A (at the trench) to point B (below the margin)separated by a distance d.

porosity are not strongly dependent, but when ΦbΦo, velocitystrongly depends on porosity and increases significantly with asmall decrease in porosity (Raymer et al., 1980). Φo is typical foreach kind of porous material. For normal compaction Ericksonand Jarrard's (1998) equation, Φo is 0.31. fsh is the shale/sand-stone content ratio. In this case, the lack of available trench coresmade us to infer fsh for both lines based on the geological settingand the seismic facies observed in the sections. Aswe stated in theprevious section, line SIS-72 shows a thick sequence of turbiditicdeposits in the trench, while line SIS-18 shows mass-wastingmaterial, probably constituted by coarser grains and rock frag-ments. Based on these observations, we considered a mid fsh,ranging between 0.30 and 0.50 for the incoming SC material inline SIS-72, and low fsh between 0.20–0.30 in SIS-18 (Table 1).Given that these estimations of fsh values can be considered assomewhat arbitrary in absence of trench sediment samplings, wehave made a test of the influence of fsh on the obtained porosity.Fig. 5 illustrate that fsh has little influence in Vp-Φ transformation.It is significant only for high pressures, but the form of the curve isidentical in the different segments. This indicates that fsh valuesslightly higher or lower to the value considered would not affectthe overall sediment behavior, the only visible effect being thedifferent dewatering fluxes.

In the second step, we calculate the amount of water confinedwithin the SC assuming that sediment pores are water-saturated.The volume of water (C) contained in a column of section (S) is

Φo Critical porosityΦi Initial porosityfsh Shale fractionC Water contentΔC Volume of water expelledΔCt Volume of water expelled per unit of time (fluid flow)Plit Lithostatic pressure (equivalent to confining pressure)Pe Effective stressPf Fluid pressure or pore pressurePhyd Hydrostatic pressureΔP Fluid overpressure or excess pore pressureλ⁎ Fluid overpressure ratiob Compressibility constantbsh Compressibility constant for shalebss Compressibility constant for sandstoneP⁎ Crushing strength

460 A. Calahorrano B. et al. / Earth and Planetary Science Letters 267 (2008) 453–467

461A. Calahorrano B. et al. / Earth and Planetary Science Letters 267 (2008) 453–467

the product of the sediment porosity in the column times the SCthickness (h) times the section:

C ¼ Uhs ð1ÞThenwe estimated the amount of water contained in sediment

pores that is likely to be expelled from the system by compactionas subduction proceeds. The differential thickness (Δh) betweena reference column of null porosity (h0) and another of porosityΦ (h) can be expressed as:

h ¼ h0 þ Dh ð2Þ

where Δh can be expressed in terms of porosity as:

Dh ¼ Uh ð3Þ

Considering Eqs. (2) and (3):

h ¼ h0= 1� Uð Þ ð4Þ

Let us then consider an element of SC with 1 km2 sectiondowngoing into subduction. Its initial state at point A, is char-acterized by hA, ΦA, and the final one, at its consecutive 100 mdistant point B, by hB, ΦB (Fig. 4). From (4) hB will be:

hB ¼ hA 1� UAð Þ= 1� UBð Þ ð5ÞAssuming that the pressure and temperature conditions within

the SC, as well as the permeability of the overriding-plate and thenature of the downgoing sediments remain basically invariablewith time at each point along the SC, sediments transported at agiven depth beneath the margin will always show the physicalproperties (Vp, Φ, etc) characteristics of that particular point. Thevolume of water that would be potentially expelled by a givenelement when moving between two points A and B can beexpressed as the difference in water content between these twopoints (ΔC=CA−CB), which is:

CA ¼ UAhA ð6aÞCB ¼ UBhA 1� UAð Þ= 1� UBð Þ ð6bÞ

and

DC ¼ hA UA � UBð Þ= 1� UBð Þ: ð7Þ

Note that Eq. (7) expresses ΔC as the difference in fluidcontent between an element of SC located in A, where it has aporosity ΦA and a volume S⁎hA, and the same element in B,where it would have a porosity ΦB and a volume S⁎hA(1–ΦA)/(1–ΦB), this is, its thickness in A corrected for the porositychange between A and B. This means that the estimation of ΔCdoes not depend on the thickness variation between the two

Fig. 6. Line SIS-72. A) 2D velocity model. B) Average Vp velocity along the SC caluncertainties. Thin dashed black line indicates the position of the deformation front.porosity and uncertainties. D) Fluid flow and water content along the subduction chb=0.052, and fluid overpressure (ΔP). G) Fluid overpressure radio (λ⁎). Color scal

points, since we do not compare different elements, but the sameelement in different locations.

Water content expressed in terms of time is:

DCt ¼ uDC=d ð8Þ

where ΔCt is the volume of water expelled per unit of time(fluid flow), u is the plate convergence rate (55 km/m.y.), and dis distance between A and B.

Finally, we estimated fluid pressures along the SC. The totalpressure supported by one point at the SC is the addition oflithostatic and tectonic pressures. Assuming that tectonic short-ening is negligiblewithin the SC, the total pressure corresponds tolithostatic pressure (Plit, i.e., confining pressure), which is com-pensated by the effective stress exerted by the solid fraction ofsubducted material (Pe), and the pressure exerted by fluid withinpores (fluid pressure or pore pressure, Pf):

Plit ¼ Pe þ Pf : ð9Þ

In our calculations, Plit includes the weights of the water andoverriding-plate sediment and basement columns. To transformvelocity into density for the slope-sediment layer, we use theGardner et al.'s (1974) relationship:

qsed ¼ 0:23Vp0:25 ð10Þ

whereas, for the upper-plate basement, we used the Nafe–Drakecurve modified by Barton (1986) for the continental rocks:

qbas ¼ 1:724þ 0:168Vp ð11Þ

which is valid for Vpb5.5 km/s.Effective stress (Pe) was estimated using a modification of

the so-called Athy's Law (Athy, 1930) proposed by Le Pichonet al. (1993). This modification consists on linking porosity withPe instead of depth, which was the variable originally used inAthy's equation:

U ¼ Uie�bPe; ð12Þ

where Φi is the initial porosity and b the compressibility constant.We consider as Φi the average velocity-derived porosity of

the 50 shallowest meters of the incoming sediment column atthe trench, and a compressibility constant, b, varying linearlybetween bsh=0.1 MPa−1 for fsh=1, and bss=0.02 MPa−1 forfsh=0 (Le Pichon et al., 1993). Interpolated b values obtainedfor lines SIS-72 and SIS-18, according to different fsh values arelisted in Table 1. Thus, we calculated Pe using Eq. (12) and thenPf using Eq. (9). Then, when Pf exceeds the hydrostaticpressure (Phyd) we got the fluid overpressure (or excess porepressure, ΔP) along the SC, ΔP=Pf−Phyd.

culated along the white dashed-line in A. The grey wide band indicates velocityThick dashed grey lines indicate limits between Zones I, II and III. C) Averageannel. E) Confining pressure F) Effective pressure (Pe) for fsh=0.40, Φ0=0.45,e represents the distance from the deformation front to each calculation point.

462 A. Calahorrano B. et al. / Earth and Planetary Science Letters 267 (2008) 453–467

463A. Calahorrano B. et al. / Earth and Planetary Science Letters 267 (2008) 453–467

Finally, a useful parameter to evaluate the drainage conditionsin the different segments of the SC is the fluid overpressure ratio,λ⁎=(Pf−Phyd)/(Plit−Phyd), which corresponds to the ratio bet-ween fluid overpressure and overburden pressure. When λ⁎=1,ΔP equals the overburden pressure, reflecting undrainedconditions or total fluid retention (the media is impermeable).Conversely, when λ⁎=0, ΔP=0 indicating optimal drainageconditions for fluid evacuation (the media is totally permeable).Table 2 shows a compilation of acronyms of the parametersconsidered for calculations.

7. Physical properties of the subduction channel

Based on Vp variations, we can divide the first tens ofkilometers of the SC along lines SIS-72 and SIS-18 in threezones (referred hereafter to as Zones I, II and III), showingcontrasting Vp-derived physical properties.

7.1. Line SIS-72

Zone I extends over the first ~9 km of the SC, and shows arapid Vp increase, from 1.8 km/s at the deformation front to2.6 km/s, ~9 km landward (Fig. 6A, B). The corresponding Vp-derived porosity drops from ~0.50 to ~0.25, and it reaches thecritical porosity at ~4 km from the deformation front (Fig. 6C).This Vp increase, andΦ reduction denote progressive compactionand subsequent dewatering of underthrust material as a responseto the increasing load of the overriding-plate. The estimatedamount of fluids expelled in Zone I, calculated by integrating thefluid flow along the whole segment, is ~12 l*yr−1 *m−2

(Fig. 6D). The gradual landward thickening and consequentload increase of the frontal prism is reflected by the steady growthof Plit from ~45 MPa at the trench, to ~70 MPa at 9 km from thedeformation front (~2 km bsf) (Fig. 6E). The porosity reductionmakes Pe to increase similarly to Plit, indicating that a significantpart of the load of the overriding-plate is transferred to the solidfraction of the underthrust material as fluids are expelled (Fig. 6F).Fluid overpressure is lower than ~8 MPa in this zone (Fig. 6F)and, consistently, λ⁎ remains inferior to ~0.5 indicating that fluiddrainage is efficient (Fig. 6G).

Zone II extends between ~9 km and ~25 km from thedeformation front. Here, the average Vp remains uniform around2.8 km/s (Fig. 6B), as well as the corresponding Φ, with valuesaround 0.28 (Fig. 6C). ThereforePe is also uniform along the entireZone II (Fig. 6F), despite the continuous increment ofPlit to ~110–120 MPa (Fig. 6E). The difference between Pe and Plit iscompensated by increasing Pf, as illustrated by the growth of ΔPfrom ~8 MPa to ~40 MPa at 25 km from the deformation front(~4.5 km bsf) (Fig. 6F). This means that the load of the overriding-plate in Zone II, is mainly supported by the pressure of pore fluidstrapped within the SC. This phenomenon agrees with the high λ⁎value of ~0.8 (Fig. 6G) that approaches to the fluid retention limit,

Fig. 7. Line SIS-18. A) 2D velocity model. B) Average velocity along the SC. The greposition of the deformation front. Thick dashed grey lines indicate limits between ZonD) Fluid flow and water content along the subduction channel. E) Confining pressufluid overpressure (ΔP). G) Fluid overpressure radio (λ⁎). Colour scale represents t

and suggests the presence of low permeable layers in this segment.Our calculations indicate that the amount of fluids that is poten-tially expelled from subducted sediments in Zone II is one order ofmagnitude smaller than in Zone I (Fig. 6D).

Zone III extends between ~25 km and ~32 km from the trench.In this zone, Vp rapidly increases from ~2.8 km/s to ~4.0 km/s(Fig. 6B), andΦ drops from ~0.25 to less than 0.10 (Fig. 6C). Thecalculated Pe grows from ~10 to ~28 MPa (from ~4.5 to 5.2 kmbsf), and the ΔP slightly decreases from ~40 MPa to 35 MParegardless of the Plit increase to ~130 MPa (Fig. 5E, F). λ⁎ dropsfrom ~0.8 to ~0.6 (Fig. 6G), suggesting enhanced fluid expulsionconditions following the rapid compaction of SC material. Theestimated amount of fluids expelled from the subducted sedimentsin this zone is ~6 l*yr−1 *m−2 (Fig. 6D).

7.2. Line SIS-18

In line SIS-18, the analysis of physical propertieswithin the SCwas done from the deformation front to ~16 km landward (~5 kmbsf) (Fig. 7). Although the average velocity is here higher than inline SIS-72, the global velocity trend evolves similarly along allzones in both lines.Whereas Vp increases along Zones I and III, itremains practically constant along Zone II (Fig. 7A, B).

The higher Vp along line SIS-18 implies that Φ must belower than in line SIS-72 (Fig. 7C). At the trench, Φi is ~0.30–0.32, indicating less pore-filling fluids entering the SC ascompared with line SIS-72, where Φi ~0.45 (Fig. 5C). In Zone I,the estimated λ⁎ is higher than in the same zone of SIS-72,suggesting less efficient drainage conditions (Fig. 7G). Theestimated fluid flow for Zone I is 4 l*yr−1 *m−2 only.

In Zone II, a constant Φ~0.25 is obtained. The estimatedfluid flow along this zone is one order of magnitude smaller thanin Zone I, indicating a low permeability zone. λ⁎ values areclose to 0.5 but increase to 0.75 over a short distance of ~3 km,pointing to strong spatial variations in permeability conditionsand drainage.

Finally, in Zone III, porosity drops to 0.10 and λ⁎ shows amarked decrease to values lower than 0.5, supporting a notableincrease of Pe arising from rapid compaction and enhanced fluidexpulsion (Fig. 7F).

8. Discussion

The quantification of physical properties of underthrustmaterialallowed differentiating three zones along the first ~30 km of theSC. Each zone is directly related to non-linear variations in Vp,Φ,andPf that suggest the presence of discrete steps of the mechanicalbehavior of underthrust material as it is buried beneath the marginwith lithostatic pressure increasing from ~0 to 130 MPa.

In the next paragraphs we compare the results obtained inZones I, II, and III in lines SIS-18 and SIS-72, and relate themwith deformation process at grain scale in order to discuss the

y wide band indicates velocity uncertainties. Thin dashed black line indicates thees I, II and III. The grey band corresponds to uncertainties. C) Average porosity.re. F) Effective pressure (Pe) for fsh=0.25 for fsh=0.25, Φ0=0.30, b=0.04, andhe distance from the deformation front to each calculation point.

464 A. Calahorrano B. et al. / Earth and Planetary Science Letters 267 (2008) 453–467

evolution of the mechanical processes governing transforma-tions of underthrust material and its influence on the stress stateof the inter-plate contact and the margin deformation.

8.1. Zone I

In both profiles, this zone is characterized by increasing Vpand reducing Φ, indicating an overall progressive consolidationof underthrust sediment, in response to the increasing load ofthe overriding-plate. At the same time, an efficient drainage of apermeable media results in a rapid expel of pore fluid, parti-cularly before the critical porosity is reached. Laboratory ex-periments indicate that at low effective pressures, b20 MPa,granular deformation is dominated by mechanisms like particlerotation and frictional slip at grain contacts, with no or very littlegrain damage (Karner et al., 2003). These mechanisms inducegranular rearrangements that increase the population of grains incontact until the system attains an optimal grain packaging,which primarily depends on its grain size and is likely to occurnear Φ0.

Pore-fluids may be expelled seaward by lateral migrationalong high-permeability stratigraphic layers (Saffer et al., 2000;Moore and Vrolijk, 1992); upward migration to the inter-platecontact and further seaward migration through the décollementzone (e.g. Moore, 1989; Saffer and Screaton, 2003), or upwardmigration across the décollement and the overriding-plate or theactive thrusts of the frontal prism (e.g. Cloos, 1984; Brown andWestbrook, 1987; Screaton and Saffer, 2005). In both lines, thehigh reflectivity of the décollement and the landward-dippingreflections of the small prism in line SIS-72 are probably ex-plained by the presence of fluids, suggesting they act as seawardfluid escape conduits.

8.2. Zone II

In this zone, Φ drops below the Φo to 0.25–0.28 (Figs. 6Cand 7C), indicating that underthrust sediment have attained itsoptimal packaging and behave now as a consolidate rock.Steady values of porosity regardless of the increasing Plit reflecta strong deceleration of sediment dewatering and consolidationrate. The growing fluid pressure that compensates the increasingload of the upper-plate, with a lower influence on effectivepressure and concomitant porosity reduction, explains thisbehavior. Grain deformation mechanisms must be similar tothose dominating at the end of Zone I, as effective pressureconditions remain practically the same. Undrained conditionsalong this zone are corroborated by high λ⁎~0.8 that indicates alost of permeability form Zones I to II. This results areconsistent with those obtained with numerical modeling ofScreaton (2006), showing a similar evolution of porosity, excesspore pressure and λ⁎ for underthrust sediments of a nearly non-accretionary margin at distances from the trench that correspondwell with those of our Zones I and II.

The low permeability of Zone II could be either related to theintrinsic nature of the underthrust sediments (e.g. Bryant et al.,1975; Screaton and Saffer, 2005), to a different physical behaviorof subducted sediments (e.g. Kimura et al., 2007), to the

mechanical characteristics of the décollement (e.g. Tobin et al.,2001), or to the overriding-plate basement acting as an imperme-able barrier. This last case seems to be less probable as Sage et al.(2006) showed high fluid over-pressures along the SC in thecentral Ecuadorian margin, regardless of a pervasive normal-faulted basement.

8.3. Zone III (and deeper)

This zone is characterized by a drop of porosity to ~0.10–0.05, together with an increase of the effective pressure, and thesubsequent decrease in fluid overpressure and λ⁎ that indicates arapid expel of over-pressured fluids out from the system. In thiscase, the system must be better drained, and the progressivelyincreasing load of the overriding-plate (Plit from ~70 to110MPa) is mainly transferred into increasing effective pressure(up to ~30 MPa). Experimental results show that for increasingPe between 30 MPa and 100 MPa, compaction also increasesresponding mainly to elastic deformation of grains (i.e. Hertziandeformation) (Terzaghi, 1925).

The over-pressured fluidsmay either escape seaward along thedécollement, or, more likely, they may migrate upward throughthe overriding-plate. In the latter case, fluids may reach thelandward-dipping faults of the overriding-plate basement, imagedin the seismic records (Fig. 2A, B), and migrate up to the seafloor.The presence of fluids would also enhance reflectivity for seismicimaging. If this is the case, seeps of fluids would be expected atthe seafloor near CDP 11000 in profile SIS-72 (Fig. 2A), as it wasobserved offshore the Pacific margin of Costa Rica (Hensen et al.,2004). It has been also described that fluid over-pressures itselfmay enhance the breakage of the permeable barrier (Sibson, 1981)and thus, the released over-pressured fluids have the potential toundermine the underside of the overriding plate by means ofhydrofracturation. This mechanism may enhance basal erosion asobserved in other margins (i.e. Muruachi and Ludwig, 1980;Ranero and von Huene, 2000; von Huene et al., 2004).

Deeper than Zone III, the granular media continuescompaction and it densifies (Karner et al., 2003). The grainpopulation reaching the failure threshold gradually increases.This process eventually leads to the macroscopic yield stagewhere grain cracking becomes pervasive (i.e., cataclasis),compactive strain rates accelerate, and the system catastrophi-cally collapses (e.g., Wong et al., 1997). Reported crushingstrengths (P⁎) are quite variable (e.g. Zoback and Byerlee,1976), and depend on the loading style and it decreases withincreasing grain size and increasing porosity (Zhang et al.,1990, Wong et al., 1997). One possibility is that the higherpressures at which the different transformational steps occur inline SIS-72 in comparison to line SIS-18 could be explained bya larger grain size of subducted sediments in SIS-18, whichagree in turn with observations suggesting turbidites andpelagic-hemipelagic deposits of the oceanic crust as potentialsource of material feeding the SC in line SIS-72, and coarsemass-wasting deposits eroded from the margin's front in lineSIS-18.

Additionally, diagenetic processes like pressure solution startto be important enhancing consolidation and lithification of the

465A. Calahorrano B. et al. / Earth and Planetary Science Letters 267 (2008) 453–467

underthrust material. These phenomena could be the nextdeformation stage, ahead of the 30 first kilometers from thedeformation front, where the aseismic–seismic transition zone islocated. An example is the Mugi tectonic Melange (Onishi andKimura, 1995; Kitamura et al., 2005) which has recorded theprocesses of underthrusting to underplating under temperatureconditions of 120–220 °C and depth of 6–7 km depth, similar tothe conditions near the updip limit of the Seismogenic Zone(Ikesawa et al., 2005; Matsumara et al., in press). In our case, asimple thermal model based on the plate geometry derived fromPSDM line SIS-72, and assuming a 34My age for the downgoingplate and thermal parameters from Marcaillou et al. (2006)predicts that temperature in Zone III of SIS-72 is close to ~80 °C(Calahorrano, 2005). Such a temperature range, associated withlithostatic pressure of 110 to 130 MPa, and effective pressure~30 MPa, corroborates our interpretation of the dominantdeformation processes affecting underthrust sediments up to6–7 km depth, and suggest that Zone III is practically precedingthe theoretical emplacement of the first diagenetic and low-grademetamorphic processes associated to the aseismic/seismic limitof the subduction megathrust.

9. Conclusion

We have quantified the along-strike and across-strikevariations of physical properties of incoming sediments alongthe first ~30 km of subduction, based on two MCS linesacquired north (SIS-18) and south (SIS-72) of the GFZ, off theerosional southern Ecuadorian margin. The analysis includesthe estimation of PSDM velocity and uncertainty within the SCand its propagation to other velocity-derived parameters such asporosity, fluid or effective pressures.

Results indicate a similar evolution of the behaviour of SCmaterial in both profiles, with increasing confining pressure. Weidentified three zones based on the non-linear variations ofphysical properties: Zone I (0–9 km from the deformation front inSIS-72, 0–5 km in SIS-18) shows a progressive increase ofseismic velocity and porosity reduction, related to sedimentcompaction and effective fluid drainage along the décollementand thrust faults of the frontal prism. Zone II (9–25 km in SIS-72,5–12 km in SIS-18) displays remarkably uniform velocity andporosity regardless of the increasing confining pressure, indicat-ing undrained conditions and thus progressively landward-in-creasing fluid over-pressureswithin the SC. Finally, Zone III (25–30 km in SIS-72, 12–16 km in SIS-18) is characterized by asudden velocity increase and porosity decrease suggesting asudden release of over-pressured fluids.

Changing physical properties estimated along Zones I, II and IIIare in agreement with experimental results for compaction andgrain deformation at different effective pressure ranges. At lowpressures, the dominant processes are granular flowand re-packing(Zone I), at mid pressures, compaction stops owing to fluid reten-tion that makes fluid overpressure to support the progressivelyincreasing margin load (Zone II), whereas at high pressures, onceover-pressured fluids are released, compaction starts to begoverned by Hertzian-like, elastic deformation (Zone III). Thismust be the dominant process until the grains attain their crushing

strength, leading to pervasive granular cataclasis and, eventually,to the collapse of the system.

Although lines SIS-72 and SIS-18 are only 60 km apart, theirSC shows different thicknesses, seismic characters and Vp thatare related to a variable nature and volume of sedimentary input.We suggest that the smaller grain size of the sediments enteringthe subduction zone in line SIS-72, as indicated by the seismiccharacter of trench sediments, may explain the higher pressuresat which the different transformations occur in this line withrespect to SIS-18.

We suggest that the sudden fluid released in Zone III mayinduce hydrofracturing favoring basal erosion, and probablyinstall the adequate conditions for the beginning of earthquakegeneration, near the updip limit of the seismogenic zone.

Acknowledgements

This work is part of the Ph. D. research (UPMC) of A.Calahorrano B., supported by a grant of the Institut de Recherchepour le Développement-IRD. A. Calahorrano has been suportedby the Juan de la Cierva Program of the Spanish Ministry ofEducation and Science during the writing of this paper.SISTEUR project was funded by the French institutes IRD,CNRS, IFREMER, UPMC. Computing and seismic processingfacilities were supported by Geosciences Azur and IFM-GEOMAR using the Large Scale Facilities of the Europeanprogram ‘Improving Human Potential’. C. R. Ranero, V. Sallarèsand A. Calahorrano B. are members of the Barcelona Center forSubsurface Imaging (Barcelona CSI) that is supported by theKALEIDOSCOPE project from REPSOL-YPF. This is acontribution of the UMR Geosciences Azur 6526, which ispart of the Observatoire de la Côte d'Azur. A. Calahorrano B. isgrateful to co-authors for fruitful discussions that help tosignificantly improve the initial manuscript and to an anon-ymous reviewer for its careful and critical reviews.

References

Agudelo, W., 2005. Imagerie sismique quantitative de la marge convergented’Equateur-Colombie: application des méthodes tomographiques aux donnéesde sismique réflexion multitrace et réfraction-réflexion grand-angle descampagnes SISTEUR et SALIERI. Univerité Pierre et Marie Curie, 203 pp.

Athy, L.F., 1930. Density, porosity and compaction sedimentary rocks. Am.Assoc. Pet. Geol. Bull. 14, 1–24.

Bangs, N.L., Westbrook, G.K., Ladd, J.W., Buhl, P., 1990. Seismic velocitiesfrom the Barbados Ridge complex: indicator of high pore pressures in anaccretionnary complex. J. Geophys. Res. 95, 8767–8782.

Bangs, N., Shipley, T.H., Moore, J.C., Moore, G.F., 1999. Fluid accumulation andchanneling along the northern Barbados Ridge dcollement thrust. J. Geophys.Res. 104, 20399–20414.

Bangs, N., Shipley, T.H., Gulik, S., Moore, G.F., Kuromoto, S., 2004. Evolutionof the Nanakai Trough décollement from the trench into the seismogeniczone: Inferences from three dimensional seismic reflection imaging. Geology32 (4), 273–276.

Barckhausen, U., Ranero, C.R., von Huene, R., Cande, S.C., Roeser, H.A., 2001.Revised tectonic boundaries in the Cocos Plate off Costa-Rica: implicationsfor the segmentation of the convergent margin and for plate tectonic models.J. Geophys. Res. 106, 19207–19220.

Bartolomé, R., Contrucci, I., Nouzé, H., Thiebot, E., Klingelhoefer, F., 2005.Using the OBS wide-angle reflexion/refraction velocities to perform pre-

466 A. Calahorrano B. et al. / Earth and Planetary Science Letters 267 (2008) 453–467

stack depth migration image of the single bubble multichannel seismic:example of the Moroccan margin. J. Appl. Geophys. 57, 107–118.

Barton, P.J., 1986. The relationship between seismic velocity and density in thecontinental crust- a useful constraint? Geophys. J. R.Astron. Soc. 87, 195–208.

Brown, K., Westbrook, G., 1987. The tectonic fabric of the Barbados Ridgeaccretionary complex. Mar. Pet. Geol. 4, 71–81.

Bryant, W.R., Hottman, W., Trabant, P., 1975. Permeability of unconsolidatedand consolodated marine sediments, Gulf of Mexico. Mar. Geotechnol. 1,1–14.

Bolton, A.J., Vannucchi, P., Clennell, M.B., Maltman, A., 2000. Microstructuraland geomechanical constraints on fluid flow at the Costa Rica convergentmargin, Ocean Drilling Program Leg 170. Proc. ODP, Sci. Results, 170:College Station, TX (Ocean Drilling Program). doi:10.2973/odp.proc.sr.170.007.2000.

Calahorrano B., A. 2005. Structure de la Marge du Golfe de Guayaquil (Equateur)et propriétés physiques du chenal de subduction, à partir des donnés desismique marine, réflexion et réfraction. Ph.D document, Univeristé Pierre etMarie Curie-Paris 6, 221p.

Cloos, M., 1984. Landward dipping reflectors in accretionary wedges: activedewatering conduits? Geology 12, 519–522.

Cloos, M., Shreve, R.L., 1988. Subduction-channel model of prism accretion,melange formation, sediment subduction, and subduction erosion at con-vergent plate margins, 1, Background and description. Pure Appl. Geophys.128, 455–500.

Clowes, R.M., Brandon, M.T., Green, A.G., Yorath, C.J., Sutherland-Brown, A.,Kanasewich, E.R., Spencer, C.S., 1987. LITHOPROBE southern VancouverIsland: Cenozoic subduction complex imaged by deep seismic reflections.Can. J. Earth Sci. 24, 31–51.

Cochrane, G.R., Moore, J.C., MacKay, M.E., Moore, G.F., 1994. Velocity andinferred porosity model of the Oregon accretionary prism from multichannelseismic reflection data: Implications on sediment dewatering and over-pressure. J. Geophys. Res. 99 (B4), 7033–7043.

Collot, J.-Y., Charvis, P., Gutscher, M.A., Operto, S., et al., 2002. Exploring theEcuador-Colombia active margin and interplate seismogenic zone. EOSTrans., Amer. Geophys. Union 83 (17), 189–190.

Collot, J.-Y., Migeon, S., Spence, G., Legonidec, Y., Marcaillou, B., Schneider,J.L., Michaud, F., Alvarado, A., Lebrun, J.F., Sosson, M., Pazmino, A.,2005. Seafloor margin map helps in understanding subduction earthquakes.EOS Trans., Amer. Geophys. Union 86 (46), 463–465.

Davis, E.E., Hyndman, R.D., 1989. Accretion and recent deformation of sedi-ments along the northern Cascadia subduction zone. Geol. Soc. Amer. Bull.101, 1153–1172.

Davis, D., Suppe, J., Dahlen, F.A., 1983. Mechanics of fold-and-thrust belts andaccretionary wedges. J. Geophys. Res. 88 (B2), 1153–1172.

Ego, F., Sébrier, M., Lavenu, A., Yepes, H., Egues, A., 1996. Quaternary state ofstress in the Northern Andes and the restraining bend model for the Ecua-dorian Andes. Tectonophysics 259, 101–116.

Erickson, S.N., Jarrard, R.D., 1998. Velocity–porosity relationships for water-saturated siliciclastic sediments. J. Geophys. Res. 103 (B12), 30385–30406.

Gardner, G.H.F., Gardner, L.W., Gregory, A.R., 1974. Formation velocity anddensity — the diagnostic basics for stratigraphic traps. Geophysics 39,770–780.

Gettemy, G.L., Tobin, H.J., 2003. Tectonic signatures in centimeter-scale velocity–porosity relationships of Costa Rica convergent margin. J. Geophys. Res. 108(B10), 2494–2505.

Guo, N., Fagin, S., 2002. Becoming effective velocity-model builders and depthimagers, Part 1 —the basics of prestack depth migration. Lead. Edge1205–1209.

Handschumacher, D.W., 1976. Post-Eocene plate tectonics of Eastern Pacific.The Geophysics of the Pacific Basin and Its Margin: American Geophys.Union Geophys. Mon., 19, pp. 799–804.

Hensen, C., Wallmann, K., et al., 2004. Fluid expulsion related to mud extrusionoff Costa Rica— a window to the subducting slab. Geology 32 (3), 201–204.

Hey, R.N., 1977. Tectonic evolution of the Cocos–Nazca spreading center.Geol. Soc. Amer. Bull. 88, 1414–1420.

Hyndman, R.D., Yorath, C.Y., Clowes, R.M., Davis, E.E., 1990. The northernCascadia subduction zone at Vancouver Island: seismic structure andtectonic. Can. J. Earth Sci. 27, 313–329.

Hyndman, R.D., Wang, K., Yamano, M., 1995. Thermal constraints on theseismogenic portion of the southwestern Japan subduction thrust. J.Geophys. Res. 100, 15373–15392.

Ikesawa, E., Kimura, G., Sato, K., Ikehara-Ohmori, K., Kitamura, Y.,Yamaguchi, A., Ujiie, K., Kato, A., Hashimoto, Y., 2005. Cataclasticinvolvement of basement basalts into plate boundary mélange: an asperitybreakage in seismogenic zone? Inference from on-land mélange in theShimanto Belt of eastern Shikoku, southwest Japan. Tectonophysics 401,217–2300.

Karner, S.L., Chester, F.M., Kronenberg, A.K., Chester, J.S., 2003. Subcriticalcompaction and yielding of granular quartz sand. Tectonophysics 377, 357–381.

Kastner, M., Elderfield, H., Martin, J.B., 1991. Fluids in convergent margins:what do we know about their composition, origin, role in diagenesis andimportance for oceanic chemical fluxes? Philos. Trans. R. Soc. Lond. Ser. A:Math. Phys. Sci. 335, 243–259.

Kimura, G., Kitamura, Y., Hashimoto, Y., Yamagushi, A., Shibata, T., Ujiie, K.,Okamoto, S., 2007. Transition of accretionary wedge structures around theup-dip limit of the seismogenic subduction zone. Earth Planet. Sci. Lett. 255,471–484.

Kitamura, Y., Sato, K., Ikesawa, E., Ikehara-Ohmori, K., Kimura, G., Kondo,H., Ujiie, K., Onishi, C.T., Kawabata, K., Hashimoto, Y., Mukoyoshi, H.,Masago, H., 2005. Melange and its seismogenic roof décollement: a plateboundary fault rock in subduction zone: an example from the Shimanto Belt,Japan. Tectonics 24.

Lallemand, S.E., Schnurle, P.S., Malavieille, J., 1994. Coulomb theory appliedto accretionary and non accretionary wedges: possible causes for tectonicerosion and/or frontal accretion. J. Geophys. Res. 99, 12,033–12,055.

Le Pichon, X., Henry, P., Lallemant, S., 1993. Accretion and erosion in subductionzones: the role of fluids. Annu. Rev. Earth Planet. Sci. 21, 307–331.

Lines, L., 1993. Ambiguity in analysis of velocity depth. Geophysics 58, 596–597.Lonsdale, P., 1976. Abyssal circulation of the Southern Pacific and some

geological implications. J. Geophys. Res. 81, 1163–1176.Lonsdale, P., 1978. Ecuadorian subduction system. AAPG Bull. 62 (12),

2454–2477.Lonsdale, P., Klitgord, K.D., 1978. Structure and tectonic history of the eastern

Panama Basin. Geol. Soc. Amer. Bull. 89, 981–999.Mascle, A., Moore, J.C., et al., 1988. Proc. ODP Initial Report 110. Ocean

Drilling Program, College Station TX, p. 63.Marcaillou, B., Spence, G., Collot, J.-Y., Wang, K., 2006. Thermal regime

from bottom simulating reflectors along the N Ecuador–S Colombiamargin: relation to margin segmentation and the XXth century greatsubduction earthquakes. J. Geophys. Res. 111, B12407. doi:10.1029/2005JB004239.

Matsumura, M., Hashimoto, Y., Kimura, G., Ohmori-Ikehara, K., Enjohji, M.,Ikesawa, E., 2003. Depth of oceanic-crust underplating in a subductionzone: Inferences from fluid-inclusion analyses of crack-seal veins.Geology 31, 1005–1008.

McIntosh, K.D., Sen, M.K., 2000. Geophysical evidence for dewatering anddeformation processes in the ODP Leg 170 area offshore Costa Rica. EarthPlanet. Sci. Lett. 178, 125–138.

Moore, J.C., 1989. Tectonics and hydrogeology of accretionnary prisms: role ofthe décollement zone. J. Struct. Geol. 11, 95–106.

Moore, G.F., Shipley, T.H., 1988. Behavior of the décollement at the toe of themiddle America trench. Geol. Rundsch. 77 (1), 275–284.

Moore, J.C., Saffer, D.M., 2001. Updip limit of the seismogenic zone beneaththe accretionnary prism of southwest Japan: an effect of diagenetic to low-grade metamorphic processes and increasing effective stress. Geology 29(2), 183–186.

Moore, J.C., Vrolijk, P., 1992. Fluids in accretionary prisms. Rev. Geophys. 30,113–135.

Moore, J.C., Shipboard Party ODP Leg156, 1995. Abnormal fluid pressures andfault-zone dilatation in the Barbados accretionary prism: Evidence fromlogging while drilling. Geology 23, 605–608.

Moore, G.F., Taira, A., Klaus, A., et al., 2001. Proc. ODP, Initial Rep.190. OceanDrilling Program, College Station TX, p. 87.

Muruachi, S., Ludwig, W.J., 1980. Crustal structures of the Japan Trench: theeffect of subduction of oceanic crust. Initial Rep. Deep Sea Drill. Proj. 56and 57, pp. 463–469.

467A. Calahorrano B. et al. / Earth and Planetary Science Letters 267 (2008) 453–467

Nur, A., Mavko, G., et al., 1998. Critical porosity: a key to relating physicalproperties to porosity in rocks. Lead. Edge 357–362.

Onishi, C., Kimura, G., 1995. Melange fabric and relative convergence insubduction zone. Tectonics 4, 1273–1289.

Ranero, C.R., von Huene, R., 2000. Subduction erosion along the MiddleAmerica convergent margin. Nature 404.

Ranero, C.R., Phipps Morgan, J., McIntosh, K., Reichert, C., 2003. Bending-related faulting and mantle serpentanization at the Middle America trench.Nature 425, 367–373.

Raymer, L.L., Hunt, E.R., et al., 1980. An improved sonic transit time-to-porosity transform. Trans. SPWLA Annu. Loggin Symp. 21st., pp. 1–13.

Ross, W.S., 1994. The velocity–depth ambiguity in seismic traveltime data.Geophysics 59, 830–843.

Saffer, D.M., Bekins, B.A., 1998. Episodic fluid flow in the Nankai accretionarycomplex: timescale, geochemistry, flow rates, and fluid budget. J. Geophys.Res. 103 (B12), 30351–33370.

Saffer, D.M., Screaton, E.J., 2003. Fluid flow at the toe of convergent margins:interpretation of sharp pore-water geochemical gradients. Earth Planet. Sci.Lett. 213, 261–270.

Saffer, D.M., Silver, E.A., et al., 2000. Inferred pore pressures at the Costa Ricasubduction zone: implications for dewatering processes. Earth Planet. Sci.Lett. 177, 193–207.

Sage, F., Collot, J.-Y., Ranero, C.R., 2006. Interplate patchiness and subduction–erosion mechanisms: evidence from depth-migrated seismic images at thecentral Ecuador convergent margin. Geology 34 (12), 997–1000. doi:10.1130/G22790A.1.

Sallarès, V., Charvis, P., 2003. Seismic constraints on the geodynamic evolutionof the Galapagos province. Earth Planet. Sci. Lett. 214, 545–559.

Sallarès, V., Ph. Charvis, E.R., Flueh, J., 2005. Bialas and the SALIERIScientific Party. Seismic structure of the Carnegie ridge and the nature of theGalápagos hotspot. Geophys. J. Int., 161 (3), 763–788. doi:10.1111/ j.1365-246 X.2005.02592.x.

Scholl, D.W., Marlow, M.S., Cooper, A.K., 1977. Sediment subduction andoffscraping at Pacific margins, in Island arcs, Deep Sea Trenches, and back-arcbasins. In: Talwani,M., Pitman,W.C. (Eds.), Am.Geophy.Union, pp. 199–210.

Screaton, E., 2006. Excess pore pressures within subducting sediments: does theproportion of accreted versus subducted sediments matter? Geophys. Res.Lett. 33, L10304. doi:10.1029/2006GL025737.

Screaton, E.J., Saffer, D., 2005. Fluid expulsion and overpressure developmentduring initial subduction at the Costa Rica convergent margin. Earth Planet.Sci. Lett. 233, 361–374.

Shipley, T.H., Stoffa, P.L., Dean, D.F., 1990. Underthrust sediments, fluidmigration paths, and mud volcanoes associated with accretionary wedge offCosta Rica: middle America trench. J. Geophys. Res. 95 (B6), 8743–8752.

Shreve, R.L., Cloos, M., 1986. Dynamics of sediment subduction, melangeformation, and prism accretion. J. Geophys. Res. 91 (B10), 10,229–10,245.

Sibson, R.H., 1981. Fluid flow accompanying faulting: field evidence andmodels. In: Simpson, D., Richards, P.G. (Eds.), Earthquake Prediction: AnInternational Review. Maurice Ewing Series, 4. American GeophysicalUnion, Washington, DC, pp. 593–603.

Stauffer, P., Bekins, B., 2001. Modeling consolidation and dewatering near thetoe of the northern Barbados accretionary complex. J. Geophys. Res. (ISSN:0148-0227) 106 (B4). doi:10.1029/2000JB900368.

Taira, A., Hill, I., Firth, J., Berner, U., Brückmann, W., Byrne, T., Chabernaud, T.,Fisher, A., Foucher, J.P., Gamo, T., Gieskes, J., Hyndman, R.D., Karig, D.E.,Kastner,M., Kato, Y., Lallemant, S., Lu, R.,Maltman, A.,Moore, G.F.,Moran,K., Olaffson, G., Owens, W., Pickering, K., Siena, F., Taylor, E., Underwood,M.B., Wilkinson, C., Yamano, M., Zhang, J., 1992. Sediment deformation andhydrogeology of the Nankai Through accretionary prism: synthesis ofshipboard results of ODP Leg 131. Earth Planet. Sci. Lett. 109, 431–450.

Terzaghi, C., 1925. Principles of soil mechanics: II. Compressive strength ofclay. Eng. News-Rec. 95, 796–800.

Thierry, P., Operto, S., Lambaré, G., 1999. Fast 2-d ray+born migration/inversion in complex media. Geophysics 64, 162–181.

Tobin, H.J., Vannucchi, P., Meschede, M., 2001. Structure, inferred mechanicalproperties, and implications for fluid transport in the décollement zone,Costa Rica convergent margin. Geology 29 (10), 907–910.

Trenkamp, R., Kellogg, J.N., Freymueller, J.T., Mora, H.P., 2002. Wide platemargin deformation southern Central America and northwestern SouthAmerica, CASA GPS observations. J. South Am. Earth Sci. 15, 157–171.

von Huene, R., Klaeschen, D., Gutscher, M., Fruehn, J., 1998. Mass and fluidflux duing accretion at the Alaskan margin. GSA Bull. 110 (4), 468–482.

von Huene, R., Ranero, C.R., Vannucchi, P., 2004. Generic model of subductionerosion. Geology 32 (10), 913–916.

Vrolijk, P., 1990. On the mechanical role of smectite in subduction zones.Geology 18, 703–707.

Westbrook, G.K., Smith, M.J., Peacock, S.M., Poulter, M.J., 1982. Extensiveunderthrusting of undeformed sediments beneath the accretionnary complexof the Lesser Antilles subduction zone. Nature 300, 625–628.

Winter, T., Avouac, J.-P., Lavenu, A., 1993. Late Quaternary kinematics of thePallatanga strike-slip fault (Central Ecuador) from topographic measure-ments of displaced morphological features. Geophys. J. Int. 115 (3),905–920.

Witt, C., Bourgois, J., Michaud, F., Ordoñez, M., Jiménez, N., Sosson, M., 2006.Development of the Gulf of Guayaquil (Ecuador) during the Quaternary asan effect of the North Andean block tectonic escape. Tectonics 25, TC3017.doi:10.1029/2004TC001723.

Wong, T.-F., David, C., Zhu, W., 1997. The transition from brittle faulting tocataclastic flow in porous sandstones: mechanical deformation. J. Geophys.Res. 102, 3009–3025.

Yilmaz, O., 2001. Seismic Data Analysis: Processing, Inversion and Interpreta-tion of Seismic Data, Vols. 1 & 2. Society of Exploration Geophysicists,Tulsa Oklahoma. 2027 pp.

Yuan, T., Spence, G.D., Hyndman, R.D., 1994. Seismic velocities and inferredporosities in the accretionary wedge sediments at the Cascadia margin. J.Geophys. Res. 99 (B3), 4413–4427.

Zhang, J., Wong, T.-F., Davis, D.M., 1990. Micromechanics of pressure-inducedgraincrushing in porous rocks. J. Geophys.Res. 95, 341–352.

Zoback, M.D., Byerlee, J.D., 1976. Effect of high-pressure deformation onpermeability of Ottawa sand. AAPG Bull. 60, 1531–1542.