Hybrid Submarine Flows Comprising Turbidity Current

Exploring the Deep Sea and Beyond themed issue

Hybrid submarine ﬂ ows comprising turbidity current and cohesive debris ﬂ ow: Deposits, theoretical and experimental analyses, and generalized models

Peter J. Talling* National Oceanography Centre, European Way, Southampton, Hampshire SO14 3ZH, UK

ABSTRACT transformations near the site of debrite depo- sured directly within such long run out fl ows. sition, and emplaced gently to avoid mixing This paucity of direct observations provides a Hybrid fl ows comprising both turbid- with surrounding seawater. The location and stark contrast to other major sediment transport ity current and submarine debris fl ow are geometry of cohesive debrites in hybrid beds processes such as rivers that have been closely a signifi cant departure from many previous are controlled strongly by seafl oor morphol- monitored in action. Understanding submarine infl uential models for submarine sediment ogy and small changes in gradient. Debrites fl ows remains a major challenge, as the only density fl ows. Hybrid beds containing cohe- occur as fringes around raised channel-levee record we have of most fl ows is the sediment sive debrite and turbidite are common in ridges, or in the central and lowest parts of deposit that they leave behind. distal depositional environments, as shown basin plains lacking such ridges. Small varia- Previous work has illustrated how long run by detailed observations from more than tions in mud fraction produce profound out distances can be achieved by turbulent sedi- 20 modern and ancient systems worldwide. changes in cohesive strength, fl ow viscosity, ment suspensions, called turbidity currents, that Hybrid fl ows, and cohesive debris fl ows more permeability, and the time taken for excess incrementally deposit layers of clean sand and generally, are best classifi ed in terms of a pore pressures to dissipate that span mul- mud. This work includes seminal contributions continuum of decreasing cohesive debris fl ow tiple orders of magnitude. Reduction in fl ow by Bill Normark and colleagues that elegantly strength. High-strength cohesive debris fl ows speed can also cause substantial increases in combined fi eld observations from the mod- tend to be clast rich and relatively thick, viscosity and yield strength in shear thinning ern seafl oor and ancient rock with quantita- and their deposit extends back to near the muddy fl uids. Small amounts of sediment can tive modeling (Bowen et al., 1984; Normark, site of original slope failure. They are typi- dampen or extinguish turbulence, especially 1989; Normark et al., 1993, 2002, 2006). Here I cally confi ned to higher gradient continental as fl ow decelerates, affecting how sediment address a second type of long run out fl ow that slopes, but may occasionally form megabeds is supported or deposited. This ensures that reaches the distal parts of submarine fans. These on basin plains, in both cases overlain by a cohesive debris fl ows and hybrid fl ows have a hybrid fl ows include both turbidity current and thin turbidite. Intermediate-strength cohe- rich variety of behaviors. mud-rich (cohesive) debris fl ow (Talling et al., sive debris fl ows typically contain clasts, 2012a), and their deposits comprise mud-rich but their deposits may be <1 or 2 m thick on INTRODUCTION debrite sand encased within turbidite clean sand low-gradient fan fringes, and are encased in and mud. This type of deposit was described turbidite sand and mud. Clasts may be far- Submarine fl ows of sediment driven by their fi rst by Wood and Smith (1958), and was noted traveled, and meter-sized clasts can be rafted excess density can run out for tens to hundreds subsequently by Van Vliet (1978), Hiscott and long distances across very low gradients if (and on occasions thousands) of kilometers, Middleton (1979, 1980), and Ricci-Lucchi and they are less dense than surrounding fl ow. sometimes across remarkably low seafl oor gra- Valmori (1980). However, only recently has Low-strength cohesive debris fl ows gener- dients of 0.1°–0.01° (Talling et al., 2012a). They it become apparent that this type of “linked” ally lack mud clasts, and as cohesive strength dominate sediment transport into many parts of debrite-turbidite bed is common in many loca- decreases further there is a transition into the deep ocean, and produce some of the most tions worldwide (Fig. 1; Haughton et al., 2003; fl uid mud layers that do not support sand. extensive and voluminous sediment accumula- Talling et al., 2004; Amy et al., 2009; Haughton Intermediate- and low-strength cohesive tions on Earth. Understanding these fl ows is et al., 2009). Hybrid fl ow deposits are the norm debrites are consistently absent in more challenging because they are remarkably dif- rather than the exception in the distal parts of proximal parts of submarine systems, where fi cult to monitor directly. The speeds of fl ows some submarine fans (Haughton et al., 2003, faster moving sediment-charged fl ows are that run out beyond the continental slope have 2009; Talling et al., 2004, 2012a, 2012b), and more likely to be turbulent. Intermediate- been measured accurately in only a few loca- they can involve very large amounts of sedi- strength debris fl ows can run out for long tions (Heezen and Ewing, 1952, 1955; Piper ment. One of the hybrid submarine fl ows that is distances on low gradients without hydro- et al., 1999; Piper and Savoye, 1993; Mulder described here transported 10 times the annual planing. Very low strength cohesive debris et al., 1997; Khripounoff et al., 2003, 2009; sediment fl ux of all of the world’s rivers com- fl ows most likely form through late-stage Vangriesheim et al., 2009; Hsu et al., 2008; bined (Talling et al., 2007a). Understanding Carter et al., 2012), and the vertical profi le of hybrid submarine fl ows is therefore important *Email: [email protected]. sediment concentration has never been mea- for determining how sediment is transported

Geosphere; June 2013; v. 9; no. 3; p. 460–488; doi:10.1130/GES00793.1; 12 ﬁ gures. Received 10 February 2012 ♦ Revision received 30 November 2012 ♦ Accepted 19 February 2013 ♦ Published online 17 April 2013

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/3/460/3343593/460.pdf by guest on 28 September 2021 Hybrid submarine sediment ﬂ ows

A Agadir Basin - Bed 5 B Marnoso-arenacea Fm. (clast-poor debrites)

150 cm 150 cm m m µ µ

200 cm mud 1 mm 250 500 100 cm 100 cm dewatering pipes

150 cm

50 cm 50 cm 100 cm

0 cm 0 cm 50 cm m m m silt µ µ µ mud 1 mm 375 500 250 0 cm Sandstone

C Mississippi Fan D Marnoso-arenacea Fm. (clast-rich debrites) Figure 1. Sedimentary logs show- 200 cm core top 100 cm Bed 3 ing different types of hybrid beds. Bed 2.5 (A) Bed 5 in the Agadir Basin off- 100 cm shore NW Africa (Talling et al., 100 cm clasts length to 320 cm 2007a). (B) Clast-poor debrites 100 cm 50 cm 30 cm in the Marnoso-arenacea For- clasts 50 cm 50 cm mation in the Italian Apennines 50 cm mud 375 500 (Talling et al., 2012b). (C) Distal 187 250 1000

lobe of the Mississippi fan in the 0 cm 0 cm 0 cm Gulf of Mexico (Talling et al., 0 cm ripples deep grooves 2010). (D) Clast-rich debrites in the Marnoso-arenacea Forma- E Jurassic & Paleocene, North Sea subsurface reservoirs F Karoo Group, South Africa 10 tion. (E) Jurassic and Paleocene 2 15 subsurface reservoir units in 0.5 2 the North Sea (Haughton et al., 8 0 1 60–100 cm

2009) (vf—very ﬁ ne; f—ﬁ ne; 60–100 cm

m—medium; c—coarse; vc— 10 6 1 very coarse). (F) Permian Karoo 0 Group in South Africa (Hodg- 4 0.5 son, 2009). (G) Dysodilic Shale 0 G Oligocene, Carpathians, Romania 5 0 100 cm in the Carpathians in Romania 1 G 0 (Sylvester and Lowe, 2004). 2 1 J 0.5 (H) Banded slurry beds in the 0 50 cm 50 cm 0 Britannia Formation, North Sea 0 0 vf f m c vc silt (Lowe and Guy, 2000; Lowe et al., clay sand

2003). (I) Megabed in the Hecho clean sandstone parallel lamination mud clasts 0 cm muddy sandstone 0 cm Group, Spanish Pyrenees (Payros dewatering pipes sandy mudstone consolidation laminae vf m vc vf m vc sand KEY et al., 1999). (J) Debrites with mud sand mudstone dish structures convolutions mud low mud content from the Boso Penin sula, Japan (Ito, 2008). I MEGA-BEDS M3 Hecho Group, Pyrenees J Boso Peninsula, Japan M3 clean sandstone debrite matrix

30 M2b sand matrix ~10% < 20 µm

20 M2a M2a mud clasts meters Britannia Formation, N. Sea Britannia Formation, M2b

H M4 10 M2a 1 m to several tens of m tens several 1 m to M2b 0 20 cm M2a carbonate or marl clasts

Geosphere, June 2013 461

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/3/460/3343593/460.pdf by guest on 28 September 2021 Talling

globally. Hybrid deposits often contain large turbidite mud VARIATIONS amounts of organic carbon, and may be a sig- A upper clean sand from H5 H4 missing nifi cant process for burying and sequestering dilute turbidity current H4 that organic carbon in deep water (Galy et al., few or no clasts 2007; Saller et al., 2008). Hybrid debrite-turbi- H3 cohesive (mud rich) debrite in debrite dite deposits are also important because they are sharp contact sharp or transitional boundary H2 a signifi cant departure from widely cited mod- banding corrugations deformed els for submarine fl ow deposits, such as those dune x-bedding of Bouma (1962), Lowe (1982), Mutti (1992), basal clean sand from dewatering and Mulder and Alexander (2001). Such gener- forerunning turbidity current planar laminated structures or (if thin and massive) H1 alized models capture how we think submarine 5 to 200 cm Typically from late stage sand settling fl ows most commonly behave at a single loca- massive & graded tion (Bouma, 1962) or evolve spatially along thin massive basal clean sand the fl ow path (Lowe, 1982, fi g. 10 therein). can pinchout in same place as debrite Generalized models are therefore needed that include hybrid fl ow deposits, and that predict how hybrid fl ows originate and evolve. Examples of down-flow bed geometries Hybrid debrite-turbidite beds are also impor- B tant because they occur in subsurface petroleum H3 reservoirs (Haughton et al., 2003, 2009), some of flow direction H5 which hold large amounts of oil and gas. Mud- H1 rich debrite sandstone has much lower permeability, and hence lower reservoir quality, than clean turbidite sandstone. Muddy debrite sand- H5 H3 stone layers act as baffl es to fl uid fl ow within flow direction H3 the reservoir (Amy et al., 2009). It is therefore H1 important to predict more accurately the location, flat basin steepening flat basin extent, and shape of debrite and turbidite sandstone layers in reservoirs containing hybrid beds, finer and ripple x-laminated (H4) in order to extract effi ciently oil and gas reserves. flow direction H3 This contribution stresses how quantitative H5 insights from theory and laboratory experiments massive & coarser (H1) play an important role in testing hypotheses for the origin of hybrid beds, and their varied fea- Figure 2. (A) Generalized graphic log for a hybrid bed at a single location, which also shows tures. Large-scale fi eld data sets from modern common variations. The different parts of the hybrid bed are labeled H1–H5 (modifi ed after systems are also important for documenting Haughton et al., 2009). (B) Three examples of the down fl ow geometry of hybrid beds, based relationships between hybrid beds and seafl oor on bed correlations in the Agadir Basin, Marnoso-arenacea Formation and Mississippi fan. gradient, and provide direct information in source areas, which are typically inferred with less certainty from ancient core or outcrop data. In a few locations, individual ancient hybrid fl ow The second aim is to provide a quantitative fl ows with increasing cohesive strength are then deposits or packages of beds can be mapped for framework for understanding hybrid fl ows that summarized, together with how these experi- long distances, providing important insights into combines observations from laboratory experimental fl ows might scale to much larger and fl ow transformation and evolution, which goes ments and theoretical analysis. When will sand often faster submarine fl ows. This theoretical beyond the vertical sequence of deposits seen at grains or mud clasts of different sizes be sup- and experimental evidence is then used to pro- one location. This contribution therefore assem- ported by cohesive debris fl ow? What deter- vide a series of general models for how hybrid bles key information from theory, laboratory mines the mud content in the matrix of the fl ows with increasing cohesive strength will experiments, modern systems, ancient outcrops, debris fl ow deposit? Small amounts of cohesive tend to evolve and deposit sediment. and subsurface cores. mud can dampen turbulence, leading to fl ow The fi nal aim is to answer a series of key ques- transformations from fully turbulent condition tions about hybrid beds. In which depositional OBJECTIVES to laminar plug fl ow. Under what conditions will settings do hybrid beds occur, and where are they such fl ow transformation occur, and how are most common? How do cohesive debris fl ows The fi rst aim of this contribution is to sum- they recorded by deposits? A wide-ranging the- originate within hybrid submarine fl ows, and why marize the fi eld observations from modern sub- oretical analysis is presented that predicts gener- are the cohesive debris fl ow deposits often absent marine fans, ancient rock outcrops, and subsur- ally how fl ows with increasing mud content and in proximal locations? Why do cohesive debrites face oil and gas reservoirs (Figs. 1 and 2). What cohesive strength will behave. This continuum occur consistently at the same level within most features are observed consistently in hybrid is illustrated using the rheology and consolida- hybrid beds? How can lateral changes sometimes beds from these disparate locations? Which fea- tion behavior of a suspension with increasing occur from turbidite-dominated to debrite-domi- tures differ, thereby defi ning different types of kaolin clay concentration. Observations from nated hybrid beds, without changes in the total hybrid bed? laboratory experiments involving sediment bed thickness? What determines the location,

462 Geosphere, June 2013

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/3/460/3343593/460.pdf by guest on 28 September 2021 Hybrid submarine sediment ﬂ ows

extent, and planform shape of cohesive debrites suggest that en masse settling, better termed 2). Higgs (2010) noted that postdepositional (in within hybrid beds? What factors control the consolidation, tends to occur from laminar plug situ) liquefaction of turbidite sand and overly- presence (or lack) of stratigraphic clustering of fl ows (Sumner et al., 2009). Sediment can be ing background (hemipelagic) mud can gener- hybrid beds? What are the implications of this supported in debris fl ows due to matrix strength, ate muddy sandstones, and that these muddy work for predicting the distribution and geom- buoyancy (reduced density contrast between sands could contain clasts of overlying back- etry of linked debrite-turbidite beds in subsur- clasts and surrounding matrix), grain to grain ground mudstone. Higgs (2010) proposed that face hydrocarbon reservoirs? A comparison is interactions, and excess pore fl uid pressures. The such liquefaction could be generated by seismic made to the previous classifi cation scheme for term cohesive debris fl ow is used here to denote shaking. Such a process may generate muddy hybrid fl ows of Haughton et al. (2009), and it is debris fl ows in which the amount of fi ne mud, sandstones in other locations; however, I concur shown how this work complements, amends, and and resulting cohesive strength of muddy fl uid, with Haughton et al. (2010), and do not think extends that classifi cation scheme. I conclude is suffi cient to support at least sand grains. This that this process generated the muddy sandstone with suggestions for further work to clarify the does not preclude the simultaneous operation intervals described here. origin and behavior of hybrid submarine fl ows. of other support processes on occasions, such as excess pore pressure. Debris fl ows in which FIELD OBSERVATIONS OF COHESIVE TERMINOLOGY USED TO DESCRIBE cohesive strength of muddy fl uid is insuffi cient DEBRITES AND HYBRID BEDS FLOW AND DEPOSIT TYPES to support sand are termed poorly cohesive, or entirely noncohesive if they contain no cohesive Field observations from a series of deposits The terminology used herein was described mud (as in Lowe, 1976). The term liquefi ed fl ow have been summarized and synthesized from in more detail by Talling et al. (2012a). The term denotes that excess pore pressures either fully or more than 20 widespread locations. This breadth turbidity current is used to denote fl ow that is mostly support sediment, such as in the experi- of observational data is needed to illustrate the fully turbulent and that deposits clean turbidite ments of Iverson et al. (2010) and Breien et al. considerable variability in cohesive debrite and sand incrementally in a layer by layer fashion. (2010). The term hybrid fl ow denotes a single hybrid bed character (Figs. 1 and 2). Laboratory experiments suggest that incremen- fl ow event that comprises both turbidity current tal deposition of clean sand tends to occur dur- and debris fl ow (Haughton, et al., 2009). Debris Flow Deposits on Modern ing turbulent (rather than laminar) fl ow condi- Larger and smaller grains tend not to segre- Continental Slopes tions (Sumner et al., 2009). Differential settling gate from a debris fl ow, unless it has very low and spatial or temporal changes in fl ow speed strength, thereby producing predominantly Very Thick Debrites lead to graded clean-sand turbidite layers unless ungraded deposits (debrites) (Sumner et al., Cohesive debris fl ows can produce very thick fl ow is steady and uniform (Kneller and Bran- 2009). Exceptions are large outsize clasts that (tens of meters) deposits that contain abundant ney, 1995) or suppressed at high concentrations. may preferentially settle, and the uppermost few clasts chaotically distributed within a homog- If outsize mud clasts are present, they occur centimeters of the debrite that may be graded enized muddy sand matrix. These very thick along discrete horizons in the deposit. due to mixing and dilution by overlying sea- debrites have a lobate or blocky morphology, High-density turbidity currents are character- water. Debrites lack sedimentary structures often with suffi cient relief to be visible in bathy- ized by high near-bed sediment concentrations formed by bedload reworking. Chaotically dis- metric mapping of the seafl oor, and can often that produce hindered settling and dampen tur- tributed mud clasts occur in some debrites, and be mapped back to the vicinity of initial slope bulence. Sediment deposition from high-density a sharp grain size break always occurs at the top failure on the modern seafl oor. Examples of turbidity currents occurs from high-concen- of the debrite interval. Debris fl ows produce a very thick cohesive debrites include those pro- tration near-bed layers that may be laminar or deposit having a thickness more closely related duced by the 1929 Grand Banks slope failure off weakly turbulent (traction carpets or laminar to the fl ow thickness, and debris fl ows can come Newfoundland (Piper et al., 1999; Mosher and shear layers; Hiscott, 1994; Sumner et al., to a halt (freeze) such that their deposits pinch Piper, 2007), the Storegga Slide offshore Nor- 2008). These near-bed layers differ from debris out abruptly in areas of low relief (Amy et al., way (Hafl idason et al., 2005), the BIG’95 slide fl ows because they are driven by the overlying 2005; Amy and Talling, 2006; Talling et al., on the Ebro continental margin (Lastras et al., fl ow (and not their own downslope weight), 2007b). Nondepositional (bypassing) fl ows that 2005), and those resulting large-scale slope fail-

and they produce planar laminated (TB) or mas- are laminar are termed debris ﬂ ow. The term ures off the northwest African margin (Masson

sive (TA) deposits, the thicknesses of which are fl uid mud (McAnally et al., 2007) is used to et al., 1993, 1997, 2010). unrelated to the overall thickness of the fl ow denote laminar fl ows that contain only mud and Glacigenic debris fl ows on trough-mouth (Kuenen , 1966a; Bannerjee, 1977; Arnott and that lack sand or larger mud clasts. fans also produce very thick (10–50 m) cohe- Hand, 1989; Sumner et al., 2008). A nondeposi- Transitional fl ow denotes the turbulence struc- sive debrites, comprising chaotically arranged tional (bypassing) fl ow that has a high near-bed tures that occur as fl ow transforms between fully clasts in a generally ungraded fi ner grained sediment concentration, but that is turbulent, is turbulent and fully laminar (plug fl ow) states matrix (King et al., 1996; Laberg and Vorren, also termed a high-density turbidity current. (Baas and Best, 2008; Baas et al., 2009; Haughton 2000; Kilfeather et al., 2010). Individual debris If sediment fallout rates are suffi ciently et al., 2009; Sumner et al., 2009; Baas et al., 2011). fl ow lobes can extend for 100–200 km and be slow, low-density turbidity currents produce 2–10 km wide, and contain 0.1–50 km3 of sedi-

ripple-scale cross-lamination (TC), overlain by DISTINGUISHING DEBRITES FROM ment. The debris ﬂ ows initiate on gradients of

ﬁ ne-scale ﬁ ne-grained planar laminae (TD), or POSTDEPOSITIONAL PROCESSES ~1°–3°, and terminate on gradients of ~0.2°– underlain by planar lamination produced by FORMING MUDDY SAND 0.5°, with multiple lobes stacked and offset

low-amplitude bedwaves (TB-2). (King et al., 1998; Laberg and Vorren, 2000; The term debris fl ow is used for fl ows that are The muddy sandstone intervals that are Elverhoi et al., 2007). The fi nal thickness of the laminar (or almost laminar) and deposit sand included in this study have features indicative of debrite deposits suggests that the debris fl ows and mud in an en masse fashion. Experiments deposition via cohesive debris fl ow (Figs. 1 and had high yield strengths of 600 to ~13,000 Pa,

Geosphere, June 2013 463

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/3/460/3343593/460.pdf by guest on 28 September 2021 Talling

assuming (1) that their density was the same as comprising conglomerate, sandstone, or mud- show that sand-sized particles were well mixed at present (~1800 kg m–3) and (2) seafl oor gradi- stone that may be 6–20 m thick (Labaume et al., within the fl ow as it moved across the area of ents of ~0.5°–2°. These very thick debris fl ows 1987; Kleverlaan, 1987). This turbidite layer sediment bypass (Talling et al., 2007a). The tend to lack any overlying turbidites, either due can sometimes extend further down the basin hybrid fl ow was therefore at least weakly turbu- to limited mixing with seawater or because the than the debris fl ow deposits (Payros et al., lent during its initial stages. turbidity current bypassed on steeper slopes and 1999). Unlike most examples of hybrid beds The debris fl ow deposit is <150 cm thick only deposited sediment beyond the termina- with thinner debrites, the debrite interval within and comprises mainly ungraded sandstone with tion of the debris fl ows. The termination of the the megabed is not typically underlain by tur- high mud matrix content (Fig. 3). It extends Canary debris fl ow in the Madeira Abyssal Plain bidite sand, although there may be local areas for 250 km and is as wide as 80 km in places is encased within graded turbidite sand (Weaver of clast-supported conglomerate (Payros et al., (Fig. 3; Talling et al., 2007a). The debrite et al., 1994), suggesting that the debris fl ow and 1999). This may be because these particularly is found in two separate low-gradient areas turbidity current reached the basin at a similar thick debris fl ows were relatively fast moving (<0.02°), and is absent across an area of slightly time, and therefore had broadly similar speeds. and outran the turbidity current, although this steeper (0.05°) slope in the central Agadir Basin is not consistent with the lack of motion within (Fig. 3; Wynn et al., 2010). There are no grain- Thinner Debrites the plug recorded by intact pectin shells in some size discontinuities or other evidence of debris Tripsanas et al. (2008) and Tripsanas and megabeds (Kleverlaan, 1987). Alternatively, fl ow bypass within the turbidite deposit in this Piper (2008) described a range of debris fl ow these debris fl ows may have completely eroded area of steeper slope (Fig. 3B). Faster velocity and other mass fl ow deposits from the con- any initial turbidite deposits. therefore most likely caused local transforma- tinental slope offshore eastern Canada, and tion of debris fl ow back into turbidity current in within minibasins in the continental slope or at Hybrid Beds and Cohesive Debrites from this area (Wynn et al., 2010). the Sigsbee Escarpment in the Gulf of Mexico. Modern Submarine Fans These include debris avalanche, coherent but Mississippi Submarine Fan, Gulf of Mexico deformed strata within slide or slumps, and Relatively thin (<2 m) cohesive debrites are This data set is important because it illus- cohesive debrites containing clasts. The debris described in three informative large-scale data trates the behavior of clast-rich cohesive debris fl ow deposits are typically no more than ~3 m sets from modern submarine fans, showing how fl ows in a well-studied submarine fan system. thick. A striking feature of this comprehensive cohesive debrite location and extent are related Debris fl ow deposits with a remarkable frond- analysis of shallow (<20 m) cores is that turbi- to seafl oor morphology. like planform shape occur at the distal fringe of dite clean sand is rare, and that most debris fl ow the most recently active lobe of the Mississippi deposits (or other types of mass fl ow deposit) Moroccan Turbidite System, Offshore fan (Fig. 4A; Nelson et al., 1992; Twichell et al., are not associated with encasing turbidite sands. Northwest Africa 1992, 1995, 2009; Talling et al., 2010). This My experience of logging more than 30 (as A hybrid fl ow deposit has been mapped for frond-like pattern is most likely due to local much as 20 m long) sediment cores from the more than 1500 km between the Agadir Basin breaching of levees, such that debris fl ow mate- continental slope offshore from the Nile delta is and the Seine and Madeira Abyssal Plains off- rial fl owed down the dip of the levees (Talling that 1–4-m-thick debrite intervals are common, shore northwest Africa, providing an unusually et al., 2010). Nelson et al. (1992), Twichell et al. but they are not associated with encasing tur- complete view of hybrid fl ow evolution (Fig. 3; (1992), and Schwab et al. (1996) concluded that bidite sand (also see Ducassou et al., 2013). This Wynn et al., 2002; Talling et al., 2007a; Frenz these deposits comprise a complex arrangement suggests that hybrid beds are relatively rare or et al., 2008; Wynn et al., 2010; Sumner et al., of thin interbedded turbidity current and debris absent on continental slopes. The reason for this 2012). This fl ow contained ~130 km3 of sedi- fl ow material. In Talling et al. (2010), we sug- is most likely that any turbidity current formed ment, and was triggered by a landslide near the gested that some intervals previously described by dilution of debris fl ow material is highly upper Agadir Canyon. The fl ow was initially as in-place turbidites are actually clasts within a mobile, and fl ows down the relatively steep con- very powerful. It eroded ~1 m of sediment in single debrite interval that is ~1–2 m thick (Fig. tinental slope without depositing turbidite sand. locations ~300 m above the canyon fl oor (Talling 4C). These clasts include boulders with diam- Therefore, debrites and turbidite sand layers are et al., 2007a; Wynn et al., 2010), and cut a spec- eters >50 cm, which are wider than the core bar- not deposited in the same geographic location. tacular fi eld of ~1-km-wide by ~10-m-deep rel (Talling et al., 2010). This difference in inter- scours at the canyon mouth (Huvenne et al., pretation highlights the issue of distinguishing Outsize Hybrid Megabeds from 2009; Macdonald et al., 2011). Canyon-mouth intervals of intact strata from large clasts within Ancient Rock Outcrops erosion provided as much as 30 km3 of mud a debrite using relatively narrow core (Haughton Outsize hybrid megabeds comprise a clast- (Huvenne et al., 2009), thereby elevating sub- et al., 2009; Jackson et al., 2009), especially rich debrite interval that is tens of meters stantially the mud content of the fl ow. The initial because elongate bedded clasts can be fl at lying. (to 250 m) thick (Fig. 1I; Johns et al., 1981; fraction of mud in the fl ow may already have The debrite is underlain by a thin (18–30 Labaume et al., 1987; Kleverlaan, 1987; Payros been considerable, as the initial slope failure cm) layer of clean massive sand (Fig. 4C). et al., 1999). Some megabeds have been mapped involved muddy continental margin sediment. Emplacement of this clean sand layer appears for 100 km, and contain 6–60 km3 or more of The flow deposited very little sediment to be closely linked to debris fl ow deposition sediment (Johns et al., 1981; Kleverlaan, 1987). beyond the canyon mouth, across 150 km of because the clean sand layer pinches out in a Clasts are supported in a muddy matrix, and can seafl oor (Fig. 3). A remarkably subtle (but four- location similar to that of the overlying debrite . be as long as 100 m. They may comprise extra- fold) slope break from 0.05° to 0.02° eventually In Talling et al. (2010), we inferred that the basinal or intrabasinal strata. Clast size is often triggered deposition from both the turbidity cur- basal clean sand most likely settled out from ungraded vertically, but there may be grading of rent and debris fl ow (Fig. 3; Talling et al., 2007a; the overlying debris fl ows, as seen in laboratory the largest clasts (Kleverlaan, 1987). The debrite Sumner et al., 2012). Similar ratios of benthic experiments involving debris fl ows with low interval is overlain sharply by a graded turbidite foraminifera in turbidite and debrite divisions strength (~5–10 Pa; Marr et al., 2001; Sumner

464 Geosphere, June 2013

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/3/460/3343593/460.pdf by guest on 28 September 2021 Hybrid submarine sediment ﬂ ows

25°W 20°W 15°W 10°W 35° N A Seine Abyssal A Plain Agadi Morocco

Agadir r Atlas Canyon Madeira Basin Mountains Abyssal Plain 30° N Sous River Canary DA Islands 200 km CDF

Figure 3. (A) Locations of the Agadir Canyon, Agadir Basin, and Seine B and Madeira Abyssal Plains; channel network between Agadir Basin 33°0′N and Madeira Abyssal Plain; Canary debris fl ow (CDF); and debris avalanches (DA) from Canary Islands. MADEIRA ! ! Path of the fl ow that deposited Bed 48 ! 5 is shown by arrows. Box indicates AGADIR area shown in C. Bathymetric con- 49 CANYON 34 57 50 32°0′N 52 tours are spaced at 500 m intervals. 33 31 (B) Map showing seafl oor gradients 30 and extent of the thin debris fl ow 29 12 deposit within Bed 5 in the Agadir 28 AGADIR Basin. The fl ow that deposited Bed 22 24 BASIN Volcanic 5 was triggered by a landslide in the core with debrite seamounts vicinity of the Agadir Canyon ca. 31°0′N core without debrite 60 ka. The fl ow eroded huge scours Slope angle at the canyon mouth, and deposited >0.25° little sediment for ~150 km beyond 0.20° 0.15° the canyon mouth. Deposition of SELVAGE 0 50 100 km 0.10° both turbidite and debrite occurred ISLANDS 0.05° 0° beyond a subtle slope break from ! ~0.05° to ~0.01° (from Wynn et al., 18°0′W 17°0′W 16°0′W 15°0′W 14°0′W 13°0′W 2010). (C) Sedimentary logs show the thickness, grain size (f—fine; 400 km 300 km 200 km 100 km 0 km m—medium; c—coarse), and sedi- 4465 C DIRECTION FLOW mentary structures within Bed 5. Key SW Turbidite grain size 57 Intervals deposited by turbidity f m c granules > 1mm current and debris fl ow are indi- sand 34 Debrite mud clasts of mud granule cated. Numbered core locations are clasts of sand shown in Figure 3 (from Talling et al., angular 2007a). 200 33 12 50 cross-laminae planar laminae contorted laminae 28 10–50 cm erosion 100 29 31 ‘Exit Ramp’

deposit thickness (cm) 49 48 52 0 grain size FLOW DIRECTION

Geosphere, June 2013 465

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/3/460/3343593/460.pdf by guest on 28 September 2021 Talling

91° 89° 87° 85° A Mississippi Delta

29° 29°

Missississippi Canyon

Slump Deposits 27° 27°

Main Fan Channel avulsion channel

25° 25°

kilometers 50 050100

91° 89° 87° 85°

85°12′0″W 85°10′0″W 85°8′0″W B 26°46′0″N 0120.5 Kilometers

Figure 4. (A) Map of the Mississippi fan with the location (box) of the side scan 26°44′0″N sonar image shown in (B). (B) Backscatter PC-42 PC-41 image of the distal part of the Mississippi

fan, showing areas of high backscatter PC-39 intensity with an intricate ﬁ ngered shape. PC-40 PC-38 Lighter hues denote higher backscatter intensity in this image. Sediment cores are 26°42′0″N shown by yellow dots, and transect of core PC-37

sites is shown by a green line. (C) Sedimen- GC-48 PC-47 tary logs along transect of cores shown by PC-46 the green line. From Talling et al. (2010).

26°40′0″N

separate fingers on side scan imagery

1.4 km 2.8 km 1.3 km 2.2 km 1.3 km 0.4 km 0.3 km C 42 41 39 38 37 48 47 46 KEY hemipelagic drape 1 metre Debrite Interval (foram rich mud) debrite interval

clean sand

homogeneous clay silty mud (debrite matrix) rich in organics Basal Sand finely laminated fine sandstone (clasts) grey clay with undeformed (flat lying) silt laminae

466 Geosphere, June 2013

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/3/460/3343593/460.pdf by guest on 28 September 2021 Hybrid submarine sediment ﬂ ows

et al., 2009). The debrite interval is overlain by seafl oor gradients (<0.2°) and debrite thickness 10–15 km across the basin. The debrite inter- hemipelagic mud, and lacks an overlying inter- suggest that these debris fl ows had low yield vals are subdivided into two types with different val of turbidite sand or mud that is seen in many strength. Some of the clast-rich debrites are downfl ow geometries. hybrid beds (Fig. 1). These debris fl ows failed to underlain directly by relatively thin (5–10 cm) Clast-rich debrites are underlain by massive produce a signifi cant turbidity current, as cores clean sand or silt intervals. Some of these basal and relatively thin clean sandstone (Fig. 5B), located just beyond the debrite termination con- clean sand or silt intervals have well-developed and pinch out abruptly over a few kilometers tain only very thin silty laminae (Fig. 4B); this planar or cross-lamination, indicating deposi- in a downfl ow direction (Fig. 6). The massive suggests that the debris fl ows were suffi ciently tion by turbidity current, most likely in the same basal sand interval pinches out at the same loca- slow moving to not undergo signifi cant mixing event as the overlying muddy debrite (Migeon tion as the debrite, suggesting that the two inter- with seawater. It is unlikely that a large turbid- et al., 2010). vals are formed by fl ows that are closely linked. ity current produced by the debris fl ow would The fi ngered high-backscatter area asso- The sand may originate from late-stage settling bypass this part of the study area, as the seafl oor ciated with debrites surrounds the elevated of sand from the debris fl ow plug, or result has a gradient of only 0.06°. channel-levee ridge (Migeon et al., 2010); from a forerunning turbidity current that is very The abundance of organic carbon in the this suggests that the fl ows responsible for the closely linked to the debris fl ow. The massive debrite matrix indicates that the debris fl ows debrites reached the distal fan mainly through sand lacks evidence of dewatering structures, did not originate from failures on the distal fan the channel. However, debrites are absent in two indicating that it did not lubricate the overly- lobes, or failures of the main channel levee, as cores from low-backscatter areas immediately ing debris fl ow by dewatering. The composition they comprise deposits with lower organic car- adjacent to the channel (Migeon et al., 2010). I of clasts and abundance of organic material in bon content. The debris fl ow matrix came from propose that debris fl ow deposition was initiated the matrix indicates that these clast-rich debris a source more than 600 km away near the Mis- by decreasing gradients as fl ows moved radially fl ows originated outside the basin plain out- sissippi Canyon (Twichell et al., 1992; Nelson down the fl anks of the channel-levee ridge. Flow crops, but had suffi ciently low strength to travel et al., 1992; Talling et al., 2010). Organic-rich adjacent to the channel was either turbidity cur- for tens of kilometers across the basin plain debris fl ow deposits are also found near the rent, or a debris fl ow that bypassed sediment (Talling et al., 2012b). junction of the lobe’s feeder channel and main across slopes >~0.2°. Clast-poor debrites contain either no clasts, fan channel, and plug the upper parts of the main or millimeter-scale mud chips. They are typi- fan channel (Nelson et al., 1992; Twichell et al., Hybrid Beds from Ancient Submarine cally underlain by relatively thick intervals 1996). It is therefore plausible that the debris Fan Outcrops of clean sand that contain clear evidence of fl ows that reached the distal fan left a trail of Studies of rock outcrops are now described deposition via turbidity current. In some loca- debrite deposits that extends back to the vicinity starting with three of the most informative locations, clast-poor debrite intervals infi ll the of the Mississippi Canyon. This, together with tions, where individual hybrid beds, or packages relief above dune crests (Fig. 5A), suggesting the large clasts within the distal fan debrites, of beds containing hybrid beds, were mapped rather gentle emplacement. These clast-poor shows that the debris fl ows did not originate over long distances. Some of the more detailed debrites gradually taper and fi ne downbasin, through fl ow transformation from initially more outcrop studies include detailed analyses of grading into graded muddy siltstone in distal dilute fl ows in which clasts were supported pri- grain size and mud content analyses for hybrid sections. They are absent in the most proximal marily by turbulence. The largest clasts must beds, or information from clast and matrix com- outcrops, where beds comprise only turbidite have been carried initially by dense fl ow in position that helps to constrain the origin of sand and mud. It is most likely that the clast- which the clasts were buoyant or supported by sediment within the fl ows. poor debris fl ows formed via fl ow transforma- matrix strength (Talling et al., 2010). tion from initial turbidity currents that deposited Miocene Marnoso-arenacea Formation in the proximal turbidites (Talling et al., 2012b). Nile Submarine Fan, Eastern the Northern Italian Apennines This fl ow transformation may have been due Mediterranean Sea This location provides especially detailed to the development of colloidal bonds between A third key data set comes from sediment information on hybrid fl ow evolution and the mud particles as the fl ow decelerated (Sumner lobes at the mouth of the most recently active planform shape of debrites within hybrid beds et al., 2009; Talling et al., 2012b, 2012c). Clast- channel-levee system in the Nile deep-sea fan (described in more detail in Talling et al., poor debris fl ows appear to have had even lower (Ducassou et al., 2008; Migeon et al., 2010; 2012b). It is the only ancient sequence where strength than clast-rich debris fl ows (Fig. 5A), Ducassou et al., 2013). This data set is important individual hybrid beds have been mapped out and they may have mixed more readily with sur- because it shows how channel-levee topography for more than 100 km (Amy and Talling, 2006; rounding seawater, causing a downfl ow dilution affects debrite distribution. A unit with high Talling et al., 2007b, 2007c, 2012b, 2012c). The and the lateral change to deposition of graded acoustic backscatter covers an area of ~100 × beds were deposited in a relatively fl at basin silts in the distal basin (Fig. 6). A single bed can 60 km, and has an abrupt fi ngered termination plain that lacked channels. Cohesive debrites contain both clast-rich debrite and clast-poor that resembles the shape of the Mississippi fan are present locally in almost all larger volume debrite at different locations across the basin. deposits (Migeon et al., 2010). Sediment cores fl ow deposits, the volume of which indicates This may result from lateral changes in debris confi rm that this high-backscatter unit com- that they were generated by slope failure. The fl ow strength within a single fl ow event (Talling prises a series of 3–6 thin (0.3–1.2 m) debris cohesive debrite intervals are always underlain et al., 2012b). fl ow deposits, as well as intervening thin turbi- by clean sand, and overlain by turbidite sand dite beds and hemipelagic mud (Migeon et al., or mud. Debrites only occur in thick beds that Lower Pleistocene Otadai Formation in

2010). Debrite intervals contain numerous mud also contain TA and TB divisions, and are always the Boso Peninsula, Japan

clasts that are chaotically distributed in a mud- absent in thin beds only comprising TC,TD, and Ito (2008) analyzed hybrid beds within a

rich matrix (Ducassou et al., 2009; Migeon TE intervals. Debrite intervals can extend for 30-m-thick interval that could be correlated et al., 2010; Ducassou et al., 2013). The low 40–80 km down the basin axis, and at least for 20 km from mid-fan channel to distal lobe

Geosphere, June 2013 467

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/3/460/3343593/460.pdf by guest on 28 September 2021 Talling

tant information on debrite distribution. The A correlated sandstone packages show lateral changes from relatively fi xed feeder channels TM [ MS2 to lobes characterized by bed amalgamation, Bed the depth of amalgamation in the lobe settings decreasing toward the fan fringe. The cohesive CS [ debrites only occur in distal lobe settings and on the eastern side of the fan fringe, suggesting that local topography may infl uence their fi nal 1 m location (Hodgson, 2009). The debrites are fi rst seen ~15 km downfan from the feeder channel exposures. The lack of debrites in the proximal lobe and feeder channels suggests that debris B overlying bed fl ows either bypassed the areas without depositing debrite, or that the debris fl ows formed by a relatively late transformation on the distal lobe from initial turbidity currents. Two types of cohesive debrites are seen (Fig. Sst-top 1F). The fi rst type of debrite contains abun- HM dant carbonaceous organic matter dispersed 1 m within a muddy sand or muddy silt matrix. The MS1 organic material is chaotically distributed, and the debrite intervals are 1–10 cm thick. Hodg- bed son (2009) suggested that the organic-rich sediment originated from shelf sequences. The CS ] organic-rich debrite is underlain by subtly nor- { mally graded clean sand that lacks structures. HM The second type of cohesive debrites contains abundant mud clasts dispersed in a muddy sand matrix. The clast-rich cohesive debrites Figure 5. Outcrop photographs of hybrid beds containing cohesive mud-rich debrite and are typically somewhat (10–50 cm) thicker, turbidite in the Miocene Marnoso-arenacea Formation (Talling et al., 2012b). (A) Bed and the clasts are not rich in organics. Hodg- in the below-Contessa section at the Castel del Priore section comprising turbidite mud son (2009) inferred that the clasts most likely (TM), mud-rich sandstone lacking clasts (MS2), and basal clean sand (CS). The mud-rich came from a source different from that of the sandstone infi lls dune crests at the top of the basal clean sandstone, which indicates slow organic-rich sediment on the continental slope emplacement of the mud-rich sandstone by a low-strength cohesive debris fl ow. Beds con- or proximal fan. Underlying basal clean sand tain mud-rich sand intervals with clasts. (B) Thin debris fl ow deposit with mud-rich sand- is typically massive, but can show lamination stone matrix that contains boulder-sized mudstone clasts to 320 cm in length. The base of in its upper part. The transition between both the bed comprises a thin layer of clean sand (CS). The debrite is overlain by a thin interval types of debrite and the underlying clean sand of rippled clean sand (Sst top) and turbidite mud. The bed is several hundred meters below is most commonly abrupt, but can be banded the Contessa Bed in the Cabelli-1 section (no. 29). MS1 refers to mud-rich debrite sand- or loaded. stone, while HM denotes hemipelagic mud. Individual cohesive debrites layers can be walked out for ~500 m, and sometimes 1 km, but they are not laterally extensive marker hori- settings using marker (ash band) horizons. His content (4%–5% fi ner than 20 μm) seen in zons, in part due to pervasive bed amalgamation study included detailed analyses of grain size, basal cleaner sand intervals. In some cases the (Hodgson, 2009). Lateral changes occur in the and is noteworthy because the debrite matrix debrite comprises the base of the bed, with a same layer between the two types of organic- has a relatively low cohesive mud content, the subtle planar laminated fabric at its base (Ito, rich and clast-rich debrite, suggesting that they importance of which is discussed herein (also 2008, fi gs. 5B and 8A therein). In other exam- have the same overall origin. The same bed can see Talling et al., 2012c). The channel and ples, the debrite is underlain by massive clean display a lateral change from being mainly clean proximal fan outcrops lack cohesive debrites, turbidite sandstone that is normally graded and turbidite sand to being mainly cohesive debrite which are found only in the intermediate sec- coarser. (Fig. 1F). The overall bed thickness does not tions. Debrites are absent in the most distal loca- change markedly across such lateral transitions, tions that comprise only thin-bedded turbidites Permian Skoorsteenberg Formation in the and the debrite thickness is locally compensated (Ito, 2008). Karoo Basin, South Africa by the turbidite sand thickness. This geometry Debrite intervals are 0.2–1.6 m thick, have Hodgson (2009) presented a detailed analy- could suggest that the depositional processes for irregular grading patterns (Fig. 1J), and con- sis of thin (0.1–1 m) hybrid beds based on out- the cohesive debrite and massive turbidite were tain abundant clasts that are 2–60 cm long. crops in which sandstone packages (as opposed not dissimilar, such that a transition from one The mud content in debrite intervals fi ner than to individual beds) have been mapped out over process to another could occur over a relatively 20 μm is 8%–14%, which exceeds the mud tens of kilometers, thereby providing impor- short lateral distance.

468 Geosphere, June 2013

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/3/460/3343593/460.pdf by guest on 28 September 2021 Hybrid submarine sediment ﬂ ows

A Clast-Rich Debrite Clast-Poor Debrite (Intermediate Strength) (Low Strength) Key turbidite mud clast-rich debrite clast-poor debrite graded silty-mud data gap data gap clean sand cross-section cross-section flow direction flow direction plan view plan view clast-rich debris flow enters outcrop area and runs out, before abruptly terminating

Clast-Poor Debrite Model 1: B Debris flow path misses proximal outcrops Key laminar X - unlikely from paleocurrent data debris in multiple beds flow

Figure 6. (A) Simpliﬁ ed sum- plan view turbulent mary of the cross-sectional turbidity and planform shape of hybrid Clast-Poor Debrite Model 2: Debris flow present at entry point, current beds containing clast-rich and but debris flow bypasses in proximal basin clast-poor cohesive debrites in √ - bypass surface (grain size break) the Marnoso-arenacea Forma- seen in proximal outcrops tion. (B) Potential models for X - would expect thin debrite lag in the origin of clast-poor debrites, proximal basin - and not seen plan view which are absent in the proxi- bypass by mal part of the hybrid bed, not- debris flow ing aspects of the ﬁ eld evidence Clast-Poor Debrite Model 3: Flow transformation forms debris flow, in favor (√) or against (X) each flow transformation triggered by erosion of mud within outcrop area of the models (from Talling turbidity current >>> debris flow et al., 2012b). X - composition of small (< 5 mm) mud clasts in debrite shows some clasts not eroded within outcrop area

X - organic carbon within debrite matrix plan view not result of erosion in outcrop area erosion of mud

Clast-Poor Debrite Model 4: Flow transformation forms debris flow, flow transformation triggered by deceleration and mud already in flow (and not erosion) turbidity current >>> debris flow √ - transition from turbidite ripples to debrite X - debris flow would need to be slow moving to avoid mixing with seawater plan view Model 4a: debris flow runs out for 10’s km turbidity current >>> debris flow (migrates) √ - transition from turbidite ripples to debrite X - location of flow transformation would have to migrate over long distances plan view Model 4b: debris flow forms locally, near to site of eventual debrite deposition

Geosphere, June 2013 469

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/3/460/3343593/460.pdf by guest on 28 September 2021 Talling

Oligocene Flysch of the East vals of sand in the Ordovician Tourelle Forma- of ~20 cm. Debrite intervals typically have abun- Carpathians, Romania tion in Canada with abundant mud matrix that dant mud clasts (to 1 m in length but more com- Sylvester and Lowe (2004) provided an contain numerous clasts, which may be as much monly a few centimeters long) and sheared sand analy sis of two outcrops that contain two types as several meters long. These cohesive debrites patches, and their muddy matrix can contain of muddy sandstone interval that are 10–50 cm lack underlying clean sand layers, and have a fl at dispersed larger sand grains or granules. Carbo- thick, which is important because of the detailed base. Jackson et al. (2009) describe hybrid beds naceous organic material is typically abundant textural data that they present (Fig. 1G). This in outcrops with inferred to be from a mid-lobe within the debrite, and both clasts and organic work illustrates that distinct types of debrite setting. Debrites were very rare in channel out- material can be fractionated toward the top or can occur in the same sequence. The fi rst type crops that comprised amalgamated thick-bedded base of the debrite. Some of the dispersed sand of muddy sandstone is relatively coarse, poorly turbidites, and absent in outcrops representing grains can be larger than the sand grains in graded and sorted, and has a matrix mud con- more distal fan fringes that were character- underlying clean sand intervals. The composi- tent of 15%–35% (Fig. 1G; Sylvester and Lowe, ized only by sheet-like thin-bedded turbidites. tion of clasts within some beds shows that the 2004). It can contain relatively large clasts that Gonzales-Bonorino and Middleton (1976) clasts are exotic, and they cannot be matched to tend to be grouped in the upper part of the deb- inferred late-stage transitions to debris fl ow for local fan fringe deposits. The debrite interval is rite interval. This type of coarser muddy debrite beds in a nonchannelized distal setting in Devo- typically underlain by a clean sandstone interval is underlain by massive or planar laminated, nian rocks in western Argentina, although they that is either massive or less commonly lami- graded, clean sandstone deposited by turbidity do not present detailed logs describing the bed nated, that commonly shows dewatering struc- current. The upper part of this clean sandstone character. Van Vliet (1978) documents that clast- tures such as pipes and dishes. In some cases, can show banding and dewatering, although rich muddy debrite occur within a fan fringe one or more stages of sand injections penetrate these features are more poorly developed than in setting in Tertiary strata outcropping along the the overlying debrite interval. In fewer cases, the Britannia Formation (Lowe and Guy, 2000), Atlantic Coast of northern Spain. The clast- the debrite is underlain by a very thin interval as discussed subsequently. The debrite interval rich debrites occur in thicker beds with a basal of clean sandstone or siltstone, or the debrite is overlain by a grain size break that separates it clean sandstone interval that is massive or con- lacks a basal sand. The boundary between the from dilute turbidity current deposits compris- tains dish structures. Clasts can be as long as basal clean sand and overlying muddy debrite ing ripple cross laminated turbidite sand. 1 m and comprise intraformational strata. Van is commonly sharp, but it can be transitional The second type of muddy debrite sandstone Vliet (1978) attributed such clast-rich debrites over a few centimeters, or display a series of is distinctly fi ner grained, better sorted, and has to erosion by initial turbidity currents within the centimeter- to decimeter-spaced laminations. 27%–37% mud matrix (Sylvester and Lowe, fan fringe. Lowe (1982, fi g. 13C therein) illus- The debrite interval is commonly overlain by a 2004). It is the only debrite deposit described trates a hybrid bed with a relatively thin inter- laminated sand or silt interval that grades into in this contribution that occurs within ripple val of graded clean sand, overlain by a cohesive turbidite mud, both of which are undeformed. cross-laminated sandstone, as all of the other debrite, in which clasts are more common near The exotic mud clasts and abundant organic

examples are not underlain by such TC divisions. the top of the debrite. Haughton et al. (2009) material suggest that some of these ﬂ ows origi- The debrite is generally clast poor, but can con- notes that hybrid beds with clast-rich muddy nated a signiﬁ cant distance updip from the fan tain occasional small mud clasts. Its grain size debrite intervals occur in the Namurian Mam Tor fringe (Haughton et al., 2003, 2009), and did distribution is similar to that of the surrounding Sandstone in England and the Macigno Forma- not result from local or basin margin failures on

TC divisions, and it is not separated from these tion and Cilento fl ysch of the Italian Apennines. the distal fan. The absence of debrites in more adjacent intervals by a grain size break. Mud-rich debrites within hybrid beds are also proximal locations is unlikely to be due only well developed in the Eocene Tyree Formation to bed amalgamation, and suggests that debris Hybrid Beds in Other Ancient in Oregon (Haughton, 2010). fl ows either bypassed through such locations, Rock Sequences or formed through fl ow transformation from Hybrid beds have been described in a series of Hybrid Beds in Subsurface Cores turbidity currents. In Haughton et al. (2009), a other ancient rock sequences representing lobe model was favored in which the debris fl ows or more distal depositional settings. Haughton Jurassic and Paleocene Submarine Fan form via fl ow transformation due to erosion of et al. (2009) and Pyles and Jennette (2009) illus- Sequences, North Sea a muddy substrate; they attributed some debrite- trate the occurrence of hybrid beds containing Haughton et al. (2003) and Haughton et al. prone packages to periods in which the sub- cohesive debrites in the Ballybunion and Inish- (2009) provided particularly detailed descrip- marine fan was out of grade, and erosion was corker outcrops of the Carboniferous Ross Sand- tions of hybrid beds within cored sequences more common in updip locations. The basal stone in western Ireland, which have a sheet-like from Jurassic (e.g., Magnus and Miller) and sandstones described in Haughton et al. (2003, geometry and represent a lobe setting. Haughton Paleocene (e.g., Forties Sandstone) subsur- 2009) often contain abundant evidence of soft et al. (2009) attributed the hybrid beds in the face reservoir units in the North Sea (Fig. 1E). sediment deformation; they therefore sug- basal Ross Formation to a period of fan initia- Hybrid beds are common in lateral and frontal gested that dewatering of the basal sand plays tion associated with incision up slope, due to a fan fringe settings in which bed amalgamation is an important part in debris fl ow run out, and this slope that was ‘out of grade,’ although the coeval rare. In some cores, >80% of the fan fringe beds hypothesis is discussed more fully herein. upslope outcrops are not described. Fan fringe or contain a muddy debrite interval (Haughton lobe deposits in the Aberystwyth Grits include et al., 2009). Debrite intervals are rare or absent Repeated Banded Intervals in beds that contain cohesive debrite intervals that in mid-fan sequences located further upfl ow Britannia-Type Slurry Beds are 0.5–1.5 m thick with particularly large clasts, that show greater bed amalgamation, and where sometimes to several meters in length (Wood and beds comprise mainly or exclusively clean sand. Several publications have provided detailed Smith, 1959; Talling et al., 2004). Hiscott and Beds with debrites range from a few centi- descriptions of slurry beds within the Aptian Middleton (1979) describe ~40-cm-thick inter- meters to >1.5 m in thickness, with an average Britannia Sandstone Member in subsurface cores

470 Geosphere, June 2013

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/3/460/3343593/460.pdf by guest on 28 September 2021 Hybrid submarine sediment ﬂ ows

from the North Sea, where the unit is a major cohesive mud particles in the near-bed bound- turbulent turbidity currents. If the sand grains gas reservoir. This example is important because ary layer. Initially, noncohesive sand and silt were supported by excess pore pressures, as it is distinctly different from most other types particles settled out from this boundary layer was the case in the laminar dense liquefi ed of hybrid beds or cohesive debrites, in which to form lighter bands. The volume concentra- fl ows described by Breien et al. (2010), then repeated banding is less well developed. The tion of mud in the boundary layer subsequently differential settling and incremental deposition slurry beds can be unusually thick (to several increased due to continued settling, and mud would occur after those excess pore pressures tens of meters) and are dominated by repeated fl occules were broken up by turbulence, leading started to dissipate. banding at a wide variety of scales (Fig. 1H; to the suppression of turbulence in the boundary Lowe and Guy, 2000; Lowe et al., 2003; Barker layer due to cohesive mud bonds, and deposition Theoretical Approach et al., 2008). Slurry beds differ signifi cantly of a dark band. The cyclic process bears some Johnson (1970) provided a theoretical rela- from debrites formed of a single interval of resemblance to the model of Stow and Bowen tionship that predicts the maximum grain diam-

muddy sandstone matrix with chaotic clasts, (1980) for laminated turbidite mud intervals. eter supported (Dmax) as a function of the sedi- τ and such clast-rich debrites also occur in the This type of cyclic depositional process may ment mixture’s yield strength ( y), gravitational ρ –3 Britannia Member (Lowe and Guy, 2000, fi g. 8 occur, but it has yet to be reproduced in fl ume constant (g), fl uid density ( f = 1000 kg m ), ρ –3 therein). Individual couplets in slurry beds typi- experiments. and particle density ( p = 2600 kg m ), such that cally comprise a pair of light and dark bands. τ ρ ρ Water escape features in the lighter bands are EXPERIMENTAL AND THEORETICAL Dmax = (8.4 y)/( p – f)g. (1) common, but truncate against the sharp base of FRAMEWORK FOR COHESIVE darker bands. Lighter bands can also be seen to DEBRITES AND HYBRID BEDS This relationship assumes spherical grains of founder into underlying darker bands. Lowe and uniform density, and has been broadly validated Guy (2000) argued that the darker bands con- I provide here a quantitative framework for by laboratory experiments (Hampton, 1975; Amy tained higher amounts of detrital mud matrix, understanding cohesive debris fl ow and hybrid et al., 2006; Sumner et al., 2009). It shows that although this was challenged by Blackbourn fl ows that combines observations from labora- sand grains with diameters of ~250–500 μm and and Thomson (2000). Banding can occur on a tory experiments and theoretical analysis. the density of quartz (~2600 kg m–3) can be sup- very wide range of scales with individual light- Experimental and theoretical analyses play an ported in muddy fl uid having a yield strength of dark couplets ranging from 50 cm to a few milli- important role in the study of submarine fl ows only 0.1–3 Pa (Fig. 7). Many of the sand-sized meters in thickness. Slurry beds comprise inter- due to the almost complete lack of direct mea- grains in these fl ows are either quartz or other vals with numerous repeated couplets, which surements from the fl ows in the deep ocean. minerals that have a density similar to quartz can be several meters thick. It is important to This discussion forms the basis for a series of (e.g., feldspars). Platy-shaped mica grains, or distinguish between such unusually thick inter- generalized models for submarine fl ows with carbonaceous organic material with much lower vals formed of numerous repeated couplets and increasing cohesive sediment strength (outlined density, would only settle through muddy fl uids debrites comprising a single muddy sandstone later herein). with signifi cantly lower strengths than those interval (Haughton et al., 2009). shown in Figure 7. Thinner intervals (<10 cm) comprising a Support of Sand by a Muddy Fluid smaller number (typically <5) of couplets can Experimental Observations occur in hybrid beds, with banding occurring Ungraded mud-rich matrix in a cohesive Annular fl ume experiments (in which cir- between the basal clean sandstone interval and debrite provides evidence that sand grains were cular fl ow is driven by paddles) have shown the overlying debrite interval. This type of thin- supported by the strength of the surrounding how a critical mud concentration in the fl ow ner banded interval was observed in subsurface muddy fl uid, such that preferential settling and is necessary to support sand (Fig. 8; Hampton, cores from the Jurassic and Paleocene sequences segregation of the larger sand grains did not 1975; Sumner et al., 2009; Baas et al., 2011), in the North Sea, and assigned to the H3 inter- occur. Such segregation settling typifi es fully and how shearing of muddy fl uid can reduce val by Haughton et al. (2009) (Fig. 2A). Lowe and Sylvester (2004, SB5 bed) also showed a <10-cm-thick interval comprising a small num- sand ber of bands below a debrite interval. However, 10 in many hybrid beds the boundary between Figure 7. Sizes of grains that basal clean sand and overlying muddy debrite will be supported by a fl uid of is relatively sharp, and in some other cases mud variable yield strength accord- content increases in a series of 2–4 steps spaced ing to the relationship proposed over centimeters to decimeters (Haughton et al., by Hampton (1975). The maxi-

2009; my own core logs from the North Sea and mum grain size supported (Dmax) 1 supported Gulf of Mexico). The thickness and number of is related to the yield strength τ repeated banded couplets seen in the Britan- ( y), gravitational constant (g), nia Formation are therefore unusual, and can ρ –3 ﬂ uid density ( f = 1000 kg m ), (Pa) Strength Yield ρ not supported exceed greatly the thickness and number of and particle density ( p = 2600 –3 τ ρ ρ bands seen in the hybrid beds described here kg m ) by Dmax = (8.4 y)/( p – f) from other locations. (see text). The grain size range of 0.1 Lowe and Guy (2000) and Lowe et al. (2003) sand sized particles is indicated. 10 100 1000 10000 proposed that the light and dark couplets in µ slurry beds formed through cyclic building up of Grain Size ( m)

Geosphere, June 2013 471

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/3/460/3343593/460.pdf by guest on 28 September 2021 Talling

10.0 to medium (63–250 mm) sand in shearing 0.18 fl ows (Fig. 8). This result was consistent with engradual masse consolidation settling of the kaolin mud contents observed to support 0.16 3 laminar plug forms debrite sand in the static settling tube experiments of 0.14 Amy et al. (2006), and consistent with theoreti- 1.0 2 cal predictions based on the equation proposed 0.12 by Johnson (1970). Sumner et al. (2009) observed that sand could 0.10 0.1 sometimes settle out from the fl ow at a late 0.08 stage, sometimes even after the initial fl ow had 0.01 sand deposition stopped moving, for kaolin concentrations of 1 0.06 between ~10.25% and 14.25% volume (Fig. 8). Sand settles out A similar late stage of larger sand grains was 0.003 0.04 to form turbidite observed by Marr et al. (2001) from debris fl ow yield strength (Pa) of kaolin (Pa) yield strength with low strength. Initial settling of a few larger 0.02 beg sand particles appears to break cohesive bonds

Kaolin concentration mud volume ins 0 0 between mud particles locally within the plug, 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 leading to further settling of sand through the Flow speed (m/s) weakened pipe area, with upward expulsion of water through the same pipes (Sumner et al., Types of Final Deposit Flow Structure Sand-deposition occurs 2009). Some sand remained within the plug. This process can produce a distinctly cleaner mud-rich debrite Laminar plug flow sand layer at the base of the deposit, overlain sand deposited by a thicker interval of muddy debrite sand that from laminar plug Trajectory of contains pipe structures. The clean basal sand debrite sand and plug develops above decelerating flow layer can be either graded or ungraded, and late sand settling turbulent base need not be strongly normally graded (Sumner from laminar plug et al., 2009). The basal sand in the Sumner et al. clean graded turbidite Turbulent flow: mud causes (2009, their web fi g. 1) experiments comprised sand deposited layer enhanced turbulence ~50% sediment coarser than ~125 μm, while by layer in turbulent flow the overlying muddy sand had ~10% sediment Turbulent flow all overlain by mud (shown in black) coarser than ~125 μm. So, not all of the larger sand grains settled into the basal sand. The Figure 8. Phase diagram showing the relationships between fl ow speed, fraction of cohesive thickness of the basal sand can depend on the mud, fl ow structure, and the type of fi nal deposit. The fi gure combines results from recirculat- thickness of the overlying debris fl ow, and the ing fl ume experiments that constrained fl ow structure (Baas et al., 2009) and annular fl ume duration of deposition if it forms incrementally experiments that documented deposit types (Sumner et al., 2009). Both sets of experiments beneath the late stages of fl ow. involved mixtures of kaolin and water, with sand added to the annular fl ume experiments (Sumner et al., 2009). Decelerating fl ows follow trajectories from right to left (gray arrows), Mud Content in Cohesive Debrite Matrix— and are initially fully turbulent. Flow deceleration can lead to development of a laminar plug. How Clean Can Debrite Matrix Sand Be? Mud content controls whether sand deposition occurs from turbulent fl ow or from laminar plug fl ow, and determines which of three different types of deposit are formed. At low mud This is an important question because mud content sand deposition occurs from turbulent fl ow and forms graded clean turbidite sand in content determines the reservoir quality of a layer by layer fashion. At high mud contents (>14.25% kaolin clay), en masse consolidation the debrite sandstone, and because it has been of a laminar plug forms ungraded mud-rich debrite sand. At intermediate concentrations suggested that laminar debris fl ows can also late-stage sand settling occurs from the laminar plug. Suffi ciently high mud fractions will deposit clean sand having a mud content that support small mud clasts, and at even higher sediment concentrations large mud clasts can resembles that of turbidites (Shanmugam become positively buoyant. and Moiola, 1995; Talling et al., 2012c). An approximate minimum volume concentration of cohesive (<20–30 μm) fi ne mud can be calcu- its strength substantially (Hampton, 1975; from experiments that had contained >4.5%– lated for a cohesive debrite plug in which the Coussot , 1995). The sand is supported by a 7% volume kaolin. Hampton (1975) noted that sand grains are supported by the strength of the network of bonds formed by surface charges even smaller amounts of stronger clays such as mud (Kuenen , 1966b). The analysis neglects between the colloidal mud particles, and the bentonite will support sand, and that shearing other support mechanisms that may help to sup- strength of these bonds is dependent on the weakens the muddy suspension signifi cantly, port the sand grains. It is assumed that the sand mud mineralogy and water chemistry, as well such that the size of sand supported after grains are closely packed, such that the muddy as mud concentration, and shear rate. Hampton shearing was typically one-half to one-third of pore fl uid comprises 25%–50% of the deposit (1975) originally concluded that just 0.6%– that supported before shearing. Sumner et al. volume at the time of deposition (Allen, 1985, 1.5% volume of kaolin would keep fi ne sand (2009) found higher concentrations of kaolin and references therein). The volume of muddy aloft, although these fi gures were extrapolated (>~14% volume) were needed to support fi ne pore fl uid will be closer to 25%–35% in more

472 Geosphere, June 2013

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/3/460/3343593/460.pdf by guest on 28 September 2021 Hybrid submarine sediment ﬂ ows

poorly sorted sand-silt mixtures (Bandini and Support of Mud Clasts by Matrix Strength Turbulence Damping by Cohesive Mud Sathiskumar, 2009). A threshold mud concen- and Flow Transformation tration of 14% by volume within the muddy Outsize mud clasts are found within many fl uid is used to illustrate the approach, based debris fl ow deposits; their density at the time Flume experiments show that relatively on experiments using kaolin (Fig. 8; Amy et al., of deposition differs signifi cantly from that of small amounts of cohesive fi ne mud (<4% vol- 2006; Sumner et al., 2009). This results in a quartz, and therefore larger mud clasts can per- ume) can dampen turbulence very effectively, minimum cohesive mud volume concentration haps be supported (Fig. 9B). Mud-clast density especially at lower fl ow velocities when shear in the entire deposit of 7% (when there is 50% is a function of the depth to which the mud has is reduced (Fig. 8; Baas and Best, 2002; Baas pore space between sand grains) to 3.5% (when been buried and compacted, before being eroded et al., 2009, 2011; Sumner et al., 2009). This there is 25% pore space between sand grains). and incorporated into the debris fl ow. Mud den- process can cause transformation from ini- Kaolin is a relatively weak clay mineral, and it is sity in the upper 10 m of sediment below the tially fully turbulent fl ow (turbidity current) to likely to be somewhat weaker than mixtures of seafl oor can range from 1350 to 1600 kg m–3 laminar plug fl ow (debris fl ow) as the speed of clay minerals found in most turbidity currents. (Flemings et al., 2006; Tripsanas et al., 2008; the fl ow decreases. Flow transformation initi- This means that the volume of clay concentra- Expedition 333 Scientists, 2011). Lower density ates in areas of lower shear in a decelerating tions in the deposit could be lower for other mud can occur in the uppermost tens of centi- fl ow, as decreasing turbulence intensity allows clay minerals, yet still support sand (Hampton, meters below the seafl oor, but such lower den- bonds to form between mud particles. Baas and 1975). If sand grains are not closely packed, and sity mud tends to act as fl uid mud (McAnally Best (2008) illustrated how a laminar plug can more widely dispersed in the plug, then the mud et al., 2007). Mud densities that are >>1600 kg progressively thicken, such that turbulence is concentration can be much higher within the m–3 can occur if the mud is buried more deeply restricted to a progressively narrower zone at the debrite. Burial will generally result in a reduc- before being exhumed (Flemings et al., 2006). bed. It is also likely that the transformation will tion in porosity, although porosity loss may be Equation 1 was used to estimate the largest begin in the slower moving tail of a fl ow. This offset by cementation. If all of the initial poros- diameter of mud clast with variable density that will favor development of laminar debris fl ow ity is lost, then the minimum volume kaolin mud could be supported in kaolin suspensions of toward the rear of the event, and debrite deposi- concentration will rise to between 14% and 7%, varying yield strength and density (Fig. 9A). It tion toward the top of the resulting deposit. The assuming that compaction only results in the is apparent that once the mud clast has a density experiments of Baas et al. (2009) and Sumner loss of water from pore spaces. that is less than that of the surrounding muddy et al. (2009) suggest that such fl ow transform In Talling et al. (2012c), we described layers fl uid, then the fl uid will be able to support very may be commonplace in submarine fl ows con- of clean sand that pinch out abruptly in the large mud clasts (Fig. 9A). Such positively taining even a small cohesive mud fraction, as Marnoso-arenacea Formation beds. These clean buoyant mud clasts could in some cases rise to every fl ow will at some point decelerate to a sand layers contain <14% (and typically <10%) the top of the debris fl ow. Mud clasts will tend to standstill. In the experiments, fl ow transforma- mud fi ner than 20 μm, as measured in scan- be buoyant in the denser and therefore stronger tion occurred at ~1.2 m/s for suspensions with ning electron microscope images (Talling et al., and more coherent debris fl ows, but large yield ~12% volume kaolin, and at speeds of ~0.2 m/s 2012c, fi g. 7 therein). This mud content is simi- strengths may prevent buoyant clasts rising to for kaolin suspensions of ~2% volume (Fig. 8; lar to that seen in the turbidite sandstone inter- the top of the debris fl ow. Baas et al., 2009; Sumner et al., 2009). Flow vals within these beds. In contrast, mud-rich debrites in these beds contain 18%–60% mud fi ner than 20 μm, as described in the preceding discussion of the Miocene Marnoso-arenacea Figure 9 (on following page). Rheology and fl ow character of a muddy suspension comprising Formation. The clean sand debris fl ows often variable amounts of kaolin. The key (bottom right) indicates the yield strength of the muddy have a distinctive swirly texture, most likely fl uid, and letters A–E correspond to fl ow types with increasing yield strength. (A) Rheology of recording pervasive liquefaction, and their suspension with variable amounts of kaolin in freshwater (from Coussot, 1995). The viscos- abrupt pinchout within this low-gradient basin ity of the muddy fl uid is strongly shear thinning such that viscosity is signifi cantly lower at plain provides evidence of debris fl ow deposi- higher rates of shear. Fluid viscosity at low shear rates (1 s–1) is used in the calculations shown tion (Talling et al., 2012c). by C to E. (B) Maximum clasts size that can be supported by the muddy fl uid and by buoy- Further work is needed to determine whether ancy (such that the clast is less dense than the muddy fl uid) for increasing amounts of kaolin. the strength of marine clay is typically only The maximum clast size is calculated according to Hampton (1975, Equation 2 therein). The slightly weaker than that of kaolin, but assum- density of mud at depths of as much as 10 m below seafl oor is shown from Integrated Ocean ing that this is the case, cohesive debrite sand- Drilling Program Leg 308 drill sites in the Gulf of Mexico (Flemings et al., 2006). (C) Mini- β stone intervals in cores and outcrop will tend to mum fl ow depth (Hc) for motion of the muddy suspension on varying seafl oor gradient ( ) τ ρ ρ β τ ρ have minimum cohesive fi ne mud volume con- calculated using Hc = y/[( f – w)g · sin ] where y is the muddy fl uid yield strength, f is the ρ –3 centrations of ~5% to ~10% (Kuenen, 1966b). muddy fl uid density, and w is the density of seawater (1020 kg m ). Colors indicate fl ow depth Amy et al. (2006, fi g. 2 therein) summarized at which the muddy fl uid becomes turbulent (from D). (D) Flow depth at which the muddy

fi eld data suggesting that cohesive debris fl ows fl uid becomes turbulent. The calculation is based on the layer averaged velocity (ULA) of a μ ρ ρ β 2 μ tend to have cohesive (<20–30 μm) fi ne mud viscous muddy fl uid with viscosity ( ) such that ULA = [( f – w)g · sin · H ]/(2 ), and the contents in excess of 10%–15% volume. Cohe- criteria of Hampton (1972) for the boundary between turbulent and laminar fl ow such that ρ 2 τ sive debrite mud fraction should generally 1000 = f · ULA/ y. (E) Speed of a 1-m-thick layer of muddy fl uid with variable mud content on ρ ρ β 2 μ exceed that of turbidity currents (and fl uidized a seafl oor gradient of 0.1°, based on ULA = [( f – w)g · sin · H ]/(2 ). The maximum speed layers; Breien et al., 2010) in which the cohe- occurs at kaolin mud volume concentrations of ~6%, and maximum fl ow speed will occur at sive fi ne mud concentration is insuffi cient to the same kaolin mud concentration on steeper slopes or for thicker fl ows. The effects of shear support sand grains. thinning on fl ow velocity are shown schematically.

Geosphere, June 2013 473

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/3/460/3343593/460.pdf by guest on 28 September 2021 Talling

clast densities that are positively buoyant B A 10000 yield strength (Pa) viscosity (Pa s–1) 10 m at low shear rate (~1 s–1) 1000 viscosity (Pa s–1) 1 m Yield Strength = 1000 Pa at high shear rate (~100 s–1) E (40% vol kaolin; 1652 kg m-3) 100 Yield Strength = 100 Pa

) 10 cm -3 –1 shear thinning at D (30% vol kaolin; 1501 kgm ) 10 higher shear viscosity rate 1 cm Yield Strength = 10 Pa of honey C (20% vol kaolin; 1344 kg m-3) 1 or molasses 1 mm Yield Strength = 1 Pa B (13% vol kaolin; 1225 kg m-3) 0.1 viscosity 100 µm Yield Strength = 0.1 Pa yield strength (Pa) yield strength (10% vol kaolin; 1178 kgm-3)

viscosity (Pa s viscosity (Pa of olive oil A µ 0.01 10 m Yield Strength = 0.01 Pa (7% vol kaolin; 1093 kgm-3) maximum grain or clast size supported µ 0.001 viscosity 1 m 1,000 -3 2,000 0 0.1 0.2 0.3 0.4 0.5 of water grain or clast density (kgm ) Quartz 2600 kg m-3 volume concentration of kaolin mud density of mud 0-10 m below sea floor from Flemings et al. (2006)

C 100 m D 100 m E 10 m 10 m D E C D A B C 1 m B 1 m A 10 cm 10 cm Threshold Flow Depth Flow Threshold Flow Turbulent for Threshold Flow Depth Flow Threshold Motion for

1 cm 1 cm 0.001 0.01 0.1 1 10 0.001 0.01 0.1 1 10 slope (°) slope (°) shear thinning E 5 1 meter thick flow on a gradient of 0.1o Key 4.5 1000 Pa yield strength (30% kaolin) high 4 Flow type in Figure 12a strength E 3.5 100 Pa yield strength (30% kaolin) intermediate 3 Flow type in Figure 12b strength D 2.5 10 Pa yield strength (20% kaolin) yield strength may start speed (m/s) 2 C Flow type in Figure 12b/c to invalidate use of Eqn. 3 low 1.5 1 Pa yield strength (13% kaolin) strength 1 B Flow type in Figure 12c 0.5 0.1 Pa yield strength (10% kaolin) 0 fluid mud A Flow type in Figure 12d 0 0.1 0.2 0.3 0.4 0.01 Pa yield strength (7% kaolin) volume concentration of kaolin mud

Figure 9.

474 Geosphere, June 2013

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/3/460/3343593/460.pdf by guest on 28 September 2021 Hybrid submarine sediment ﬂ ows

transformation will tend to increase the viscos- and may not lubricate the trailing body of the Mixing and Dilution by Seawater ity of the fl uid through stronger mud bonds, and fl ow. Greater lubrication of the head can lead to this may cause further deceleration, producing a detachment of faster moving outrunner blocks Rates of mixing between debris fl ow and positive feedback. from the main debris fl ow. Hydroplaning may the surrounding seawater or turbidity current protect underlying deposits from erosion, but are possibly very important controls on sub- Flow States and Corresponding Deposit Types only at the front of the fl ow. marine fl ows (Kuenen, 1951; Hampton, 1972; Sumner et al. (2009) showed how depo- Mohrig et al., 1998; Marr et al., 2001; Moh- sition of turbidite or cohesive debrite sand Basal Shear Wetting: Mixing of the rig and Marr, 2003; Ilstad et al., 2004; Felix is linked to fl ow state (turbulent or laminar) Overridden Water Back into the Debris Flow and Peakall, 2006). Mixing and dilution can during sand deposition. Turbidite sand tended Experimental and theoretical work by decrease the strength and viscosity of the debris to be deposited under turbulent fl ow condi- Ilstad et al. (2004) suggests that water over- fl ow in a nonlinear fashion, and determine the tions before transformation to laminar fl ow ridden at the head will often be mixed back rate at which associated turbidity currents are occurred as fl ow decelerated (Fig. 8). Debrite into the body of the debris fl ow, due to high produced. The rate of mixing can determine the sand was deposited from laminar fl ow, if the shear rates at the bottom of the fl ow. Ilstad duration and run out distance of a debris fl ow, if transformation from turbulent to laminar fl ow et al. (2004) termed this process shear wetting, the debris fl ow is transformed completely into occurred at higher fl ow velocities before the and it would tend to produce a basal sediment turbidity current. sand started to settle out. This resulted in the layer with high water content, rather than a generation of a laminar plug fl ow from which basal layer comprising only seawater. Such a Processes of Mixing at the Head muddy debrite sand was eventually deposited basal water-rich sediment mixture could still Shear stresses between the debris fl ow and en masse. At intermediate concentrations, act to lubricate a debris fl ow. It may be even surrounding water, and erosion rates, are higher sand settled out from the laminar plug at a more important in submarine debris fl ows that at the head than above the body of the debris late stage, sometimes even after fl ow stopped tend to be faster moving than their laboratory fl ow. The head can be eroded by detachment of (Sumner et al., 2009; see also discussion of counterparts. single grains at low shear stresses that exceed Experimental Observations). This associa- a critical value. This critical value of the shear tion between the type of sand deposited and stress (typically 0.5–2 Pa) can be orders of mag- the fl ow state results from near coincidence Low-Strength Layers of Mud at the Seafl oor nitude lower than the sediment mixture’s yield between the cohesive mud content necessary The uppermost ~20 cm of the modern sea- strength, so erosion and mixing can initiate to support sand and the cohesive mud content fl oor can often have very low strength (soupy well before the material is internally deformed. needed to suppress turbulence in the experi- when cored) mud. This is why box corers As shear stresses increase at the head, intact ments. It remains to be seen whether such a (rather than piston or gravity) corers are typi- chunks of material, and then a discrete layer coincidence characterizes a wider range of cally used to sample the sediment-water inter- of sediment, may be fl ung backward toward marine mud compositions. face. It is possible that this type of low-strength the body. mud layer can also lubricate a debris fl ow that Hydroplaning passes across it, in a manner similar to shear Processes of Mixing along the Body wetting. Laboratory studies have tended to focus Laboratory experiments have shown how a on mixing at the head, as the head dominates layer of water may be injected under the head of Depositional Record of Hydroplaning smaller volume experiments. The body will a submarine debris fl ow, if the dynamic pressure and Shear Wetting? comprise a far greater proportion of most sub- developed at the nose of the fl ow exceeds the Outrunner blocks have been observed that marine debris fl ows, and may play a greater role downward directed weight of debris (Mohrig most likely result from basal lubrication of in mixing and dilution (Talling et al., 2002). Ero- et al., 1998; Harbitz et al. 2003; De Blasio et al., debris fl ows (Harbitz et al., 2003; De Blasio sion of the body will also begin through detach- 2004). Hydroplaning will therefore character- et al., 2004). However, in many cases outrun- ment of individual grains, and then proceed to ize faster debris fl ows, or thinner debris fl ows ner blocks are not observed, and the lobate erosion of chunks as shear stresses increase. As that have lower density. A further condition for submarine debris fl ow deposits resemble the with the head, experiments suggest that mixing hydroplaning is that the debris is suffi ciently deposits of terrestrial debris fl ows that did not becomes much more effi cient once the body impermeable to prevent rapid dissipation of hydroplane (e.g., Laberg and Vorren, 2000; becomes turbulent, and mixing can also occur the overridden water. Water overridden at Talling et al., 2010). Shear wetting would gen- through breaking waves along the upper surface the debris fl ow head will tend to lubricate the erate homogenized layers of fl ow with lower of the body (Felix and Peakall, 2006). debris fl ow. It has been proposed that hydro- sediment concentrations than the overlying planing is one explanation for the long run out debris fl ow. Erosion of seafl oor sediment can Potential Bifurcation in Mixing and of submarine debris fl ows across low-gradient also generate basal layers in the fl ow, such as Flow Behavior seafl oor (Mohrig et al., 1998; De Blasio et al., the muddy sand layer described by Gee et al. A reduction in sediment density, especially 2004). The injection of water during hydroplan- (1999) beneath the Canary debris fl ow. Layers through turbulent mixing of the fl ow interior, ing will only penetrate for a limited distance of massive clean sand are observed commonly will decrease the strength and viscosity of a sedi- under the head of a hydroplaning debris fl ow, below debrites, but it is diffi cult to envisage how ment mixture in a strongly nonlinear fashion. where it will tend to be shear mixed back into such clean sand layers could record basal shear This can promote more rapid mixing, and the the body of the fl ow. In the laboratory experi- wetting, or seafl oor erosion. Shear mixing with reduced viscosity will tend to increase the speed ments of Mohrig et al. (1998), this distance is either the muddy seafl oor or the muddy debris of the fl ow. This positive feedback may produce a few tens of centimeters. Hydroplaning there- fl ow would be expected to produce a muddy a bifurcation in fl ow behavior, such that some fore only lubricates the very front of the fl ow, sand layer, not a clean sand layer. fl ows undergo much more effective mixing and

Geosphere, June 2013 475

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/3/460/3343593/460.pdf by guest on 28 September 2021 Talling

ρ ρ Φ 2 μ Φ dilution than other fl ows. Flow behavior will be ULA = [( df – w)g · sin · H ]/(2 ). (3) thickness (H), and seafl oor gradient ( ), such ρ ρ Φ heavily dependent on whether the head and espe- that W = ( df – w) · H · cos . The condition for cially the body of a fl ow become turbulent. Equation 3 assumes laminar fl ow and no fric- hydroplaning (Mohrig et al., 1998) is therefore tional retardation due to mixing along the upper Φ ρ ρ ρ 0.5 Theoretical Analysis of Flow Dynamics as surface of the debris fl ow. Previous work has UF > 0.35 {[g · cos ( df – w) · H]/ w} . (5) Sediment Strength Increases commonly used an alternative Chezy-type equation to estimate the speed of turbidity cur- The analysis based on kaolin rheology sug- A theoretical analysis explores how sub- rents (Bowen et al., 1984), but such an approach gests that hydroplaning could be common, and marine fl ows behave as strength of the sediment assumes that fl ow viscosity does not play an that a suffi ciently thick high-strength mixture mixture forming the fl ow increases; this analy- important role in determining fl ow speed. of kaolin will nearly always hydroplane, if it is sis extends on work by Hampton (1975), His- A plot of fl ow speed against volume concen- able to move (Fig. 10). A fl ow (or block) that is cott and Middleton (1979), Locat et al., (1996), tration of kaolin shows that increasing viscos- hydroplaning may travel beyond the position at Schwab et al. (1996), and Talling et al. (2002, ity can become important at relatively modest which a non-hydroplaning fl ow will stop. Faster 2010). Flow evolution is illustrated by a series mud concentrations. For example, increasing moving fl ows that hydroplane may also be prone of plots based on the rheology of a kaolin sus- viscosity causes a decrease in fl ow speeds once to becoming turbulent, and this may be accen- pension (Fig. 9; rheology from Coussot, 1995). the volume concentration of kaolin in the sus- tuated by decreases in fl ow viscosity at higher pension exceeds ~6% for a 1-m-thick fl ow on a shear rates (Fig. 9E). Laboratory experiments Debris Flow Thickness Necessary for Motion gradient of 0.1° (Fig. 9E). Direct measurements have mainly focused on hydroplaning of rela- on Different Slope Gradients of hyperconcentrated fl ow speeds in the Yellow tively thin and laminar fl ows. Turbulence may A relationship exists between the minimum River in China similarly suggest that viscous cause mixing of overridden seawater back into

fl ow thickness (Hc) needed for continued motion forces become important at relatively low sedi- the main body of the fl ow in many situations. τ of debris fl ow with a certain yield strength ( y) ment concentrations, such that maximum fl ow The analysis predicts that moving fl ows with ρ and density ( df), across seafl oor with a gradient speeds are reached at ~5% sediment volume uniform thickness and composition will undergo of Φ (in degrees), such that concentration (Van Maren et al., 2009). a transition from turbulent to laminar fl ow as seafl oor gradients decline. A more enigmatic τ ρ ρ Φ Hc = y/[g( df – w) sin ], (2) Conditions for Turbulent or Laminar Flow prediction is that thicker fl ows that are turbu lent Flow depth at which the muddy fl uid becomes sometimes come to a halt without a transition ρ where w is the density of surrounding sea- turbulent can be calculated assuming the rheol- to laminar fl ow; this prediction appears to be water. The threshold thickness for fl ow motion ogy of a kaolin-water mixture shown in Figure un realistic, unless there is an abrupt transition is shown for varying seafl oor gradients and 9d, using the criteria of Hampton (1972): from turbulent fl ow to no motion. yield strength for a kaolin-water mixture (Fig. ρ 2 τ 9C). The yield strength is that in the basal part 1000 = df · U / y. (4) Rates of Shear Mixing of the fl ow, so that it is assumed that the debris Here I explore whether mixing and dilu- fl ow is not hydroplaning or otherwise lubricated The calculations use fl ow viscosity at low shear tion by surrounding seawater prevents long by shear wetting, which would increase the run rates (Fig. 9A; red dots), and fl ow viscosity will run out of low-strength debris fl ows. Labora- out of thinner fl ows onto lower slopes. It also tend to decrease at higher fl ow speeds and shear tory experiments illustrate how relatively small assumed that the debris fl ow is not continuing to rates (Fig. 9E). This means that turbulent fl ow changes in debris fl ow sediment concentration move due to momentum inherited from further may occur on somewhat lower gradients and for (e.g., 2%–3% volume concentration of kaolin; upslope (fl ow is steady), and that a stiff debris thinner fl ows than those shown in Figure 9D. Hampton, 1972) and yield strength can lead to fl ow is not being pushed from behind by faster The calculations suggest that a thin (~2 m) disproportionately large changes in the rate of moving parts of the fl ow. Repeated en masse fl ow with yield strength of <~10 Pa, on a gra- mixing and the rate of turbidity current gen- deposition from debris fl ows in multiple pulses dient of <0.05°, is likely to be at least weakly eration. Dilution of the debris fl ow can reduce can produce a deposit that is thicker than the turbulent. Thicker debris fl ows (>5 m) with its viscosity substantially, potentially leading original debris fl ow (Major and Iverson, 1999). greater yield strengths (>100 Pa) are also likely to fl ow acceleration and therefore even higher Postdepositional compaction of the initial debris to be turbulent on steeper gradients of >~0.1°. mixing rates. fl ow deposit reduces its thickness to less than This analysis suggests that the (often clast rich) However, mixing rates are perhaps the most that of the original fl ow; this must be considered sediment mixtures responsible for depositing problematic aspect of submarine fl ows to model when using deposit thickness to estimate debris cohesive debrites distally may often be at least theoretically (Talling et al., 2002). Parchure and fl ow thickness. However, the analysis shows weakly turbulent on steeper slopes closer to Mehta (1985) proposed that the rate of erosion that thin (<2 m) debrites found on low seafl oor the source. (E, in kg m–2 s–1) increased nonlinearly with gradients (<0.1°) are likely to have been depos- the excess shear stress undergone by the debris τ τ ited by debris fl ows with relatively low strengths Conditions for Hydroplaning ( – s), such that to ~10 Pa. Hydroplaning occurs when the dynamic pressure at the front of the debris fl ow exceeds E = β εα(τ – τs)0.5, (6) Flow Speed as a Function of the downward weight exerted on the overrid- α β ε Sediment Concentration den water. The frontal dynamic pressure (Pf) is where and are empirical constants and is ρ The layer-average speed of a viscous muddy related to the density of seawater ( w) and the the rate of surface erosion (m/s). Amos et al. μ ρ fl uid (ULA) with viscosity ( ), density ( df), and square of frontal fl ow speed (UF), such that Pf = (1996) suggested a linear relationship, such that Φ ρ 2 thickness (H), on a seafl oor gradient of , can ½ · w · UF. The weight of the sediment (W) ρ τ τ τ be estimated using is proportional to the debris density ( df), fl ow E = M[( – s)/ s], (7)

476 Geosphere, June 2013

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/3/460/3343593/460.pdf by guest on 28 September 2021 Hybrid submarine sediment ﬂ ows

A B layers with low yield strength. This approach 100 m 100 m is based on a bulk Richardson number (Ri ), 1000 Pa 100 Pa * expressed as a function of the shear velocity ρ ρ 10 m 10 m (u*), density contrast ( df – f), and thickness of the upper well-mixed turbulent layer (T), such that 1 m 1 m ρ ρ ρ 2 Ri* = g · T [( df – f)/ f]/(u*) . (8)

10 cm 10 cm Threshold Flow Depth Flow Threshold Flow Turbulent for Depth Flow Threshold Flow Turbulent for A substantial body of experimental work has shown how the mixing rate is related to bulk 1 cm 1 cm Richardson number (e.g., Kranenburg and 0.001 0.01 0.1 1 10 0.001 0.01 0.1 1 10 Winterwerp, 1997). However, this approach slope (°) slope (°) cannot be applied easily to submarine debris ﬂ ows, as it is unclear what value to use for the thickness (T) of the overlying well-mixed layer C 100 m D 100 m (Talling et al., 2002). Further work is needed 10 Pa 1 Pa to analyze the role of the upper layer thickness that tends to vary only over a narrow range 10 m 10 m in experiments, and its physical signiﬁ cance, if this approach is to be applied to submarine

1 m 1 m debris ﬂ ows. Flume experiments illustrate in how low- strength (<5 Pa) muddy suspensions mix 10 cm 10 cm

Threshold Flow Depth Flow Threshold Flow Turbulent for Depth Flow Threshold Flow Turbulent for rather easily and erode rapidly. Winterwerp and Kranenburg (1997) found that a suspension containing 2.5% kaolin was eroded at a 1 cm 1 cm rate of ~5 mm/s by an overlying fl ow traveling 0.001 0.01 0.1 1 10 0.001 0.01 0.1 1 10 at ~50 cm/s. A 1 m thickness of this material slope (°) slope (°) would be rapidly eroded in ~3 min. A suspension containing 7.5% volume concentration E 100 m kaolin was eroded at a rate of 0.2 mm by simi- 0.1 Pa yield strength 0.1 Pa lar fl ow speeds, such that 1 m of erosion would no motion take ~1.5 h. Winterwerp and Kranenburg (1997) 10 m allowed the kaolin suspension to consolidate for turbulent flow several hours, and estimated that it had a yield 1 m laminar flow strength of ~4 Pa; the rheology of their kaolin suspensions may span the boundary from fl uid threshold flow mud to very low strength debris fl ows that can 10 cm depth for motion Threshold Flow Depth Flow Threshold Flow Turbulent for support sand (fl ow types A and B in Fig. 9). The results of Winterwerp and Kranenburg (1997) threshold flow support qualitative observations from other 1 cm depth to hydroplane annular fl ume experiments in which mud sus- 0.001 0.01 0.1 1 10 pensions with yield strengths of <10 Pa mixed slope (°) threshold flow depth for turbulence rapidly with overlying water at fl ow speeds of a few tens of centimeters per second (Esther Figure 10. Plots of threshold values of fl ow thickness and seafl oor gradient that cause hydro- Sumner, 2010, personal commun.). This obser- planing, allow motion, and produce a transition from turbulent to laminar fl ow, for sedi- vation is important because it suggests that thin ment mixtures with variable yield strength. See text for details of equations used. (~1 m), low-strength (<1 Pa) debris fl ows could not travel for long distances at speeds of even a few tens of centimeters per second without M is an empirically defined constant. In mixtures with yield strengths of ~10 Pa could mixing almost completely with the surround- Talling et al. (2002), we showed how Equation be eroded to depths of ~50 m while fl owing ing seawater. Calculations of fl ow speeds for 7 could be used to predict the rate of mixing at these speeds over the same distance. How- such low volume concentration kaolin mixtures and erosion for stronger or weaker sediment ever, these equations are not well validated suggest that speeds of 1 m/s would often be mixtures. Strong mixtures with yield strength by experiments that have generally not simu- exceeded, even on gradients of just 0.1° (Fig. of ~5000 Pa underwent relatively little erosion; lated erosion via detached chunks at higher 9E). This suggests that very low strength (clast erosion to a depth of less than a few meters fl ow speeds. poor) cohesive debris fl ows tend not to fl ow for occurred for fl ow speeds <10 m/s across a dis- An alternative approach to mixing across a long distances, but form at a late stage from fl ow tance of 100 km. In contrast, weak sediment density interface has been used for fl uid mud transformation.

Geosphere, June 2013 477

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/3/460/3343593/460.pdf by guest on 28 September 2021 Talling

Time Taken for Excess Pore C = 10–7 108 Pressures to Dissipate 1 year C = 10–6 Changes in excess pore pressure can pro- 107 C = 10–5 foundly change cohesive debris fl ow motion 6 10 C –4 (Iverson and Vallence, 2001; Iverson et al., 1 week = 10 5 2010), and indeed in fl ows in which the excess 1 day 10 pore fl uid pressure rather than cohesive matrix 104 strength supports sand grains. Compaction of 1 hour 1000 the debris during motion would cause increases in the debris fl ow’s strength and viscosity. 100 Dissipation of excess pore pressures after the 10 debris fl ow has stopped can lead to foundering 1

of overlying layers, and in situ soft sediment decay to pressure pore basal excess time for 1/e (37%) of initial value to deformation. 0.1 0.01 0.1 1 10 The time taken for the basal pore pressure to debris flow thickness (m) dissipate to 1/e of its original value (tpp) is deter- mined by the coeffi cient of consolidation (C), Figure 11. Time taken for basal excess pore pressure to decay to 1/e which is sometimes termed the hydraulic diffu- (37%) of its initial value as a function of debris fl ow thickness, for sivity (Iverson et al., 2010), and fl ow thickness different values of the coeffi cient of consolidation (C) (after Iver- (H) such that son et al., 2010). The analysis assumes that larger grains do not preferentially settle out from the sediment-water mixture, and that 2 tpp = 2C/H . (9) en masse consolidation occurs.

The coeffi cient of consolidation depends upon the permeability and stiffness of the material, and can therefore vary many orders of magni- amounts of cohesive mud (Talling et al., 2012a). Mohrig and Marr, 2003; Ilstadt et al., 2004; tude. For example, the addition of ~5% volume Small volume fractions (<<1%) of noncohesive Breien et al., 2010). It can be useful to group mud fi ner than 64 mm to an initial mixture of grains can also dampen turbulence (Wright and experiments according to the coherency of sand and gravel reduced the coeffi cient of con- Parker, 2004; Cantero et al., 2012). This sug- the debris fl ow (Marr et al., 2001). Coherency solidation from 10–4 to 10–7 in the experiments gests that transitions from turbulent to laminar expresses the relative magnitudes of the sedi- of Iverson et al. (2010). As the time taken for fl ow may be relatively common as submarine ment’s yield strength and dynamic pressures it pore pressure to dissipate is proportional to the fl ows decelerate. undergoes due to shear with surrounding water coeffi cient of consolidation, small increases in and seafl oor. Strongly coherent fl ows tend to mud content profoundly change the period over Shear Thinning and Potential Bifurcation have higher yield strengths and/or travel at which excess pore pressures remain (Fig. 11). in Flow Behavior lower speeds, as the shear forces tend to scale Mud contents necessary to support sand are with the square of fl ow speed. likely to result in very low coeffi cients of con- A bifurcation in fl ow behavior may result solidation (<10–7); this means that little excess from the shear thinning properties of many Strongly Coherent Experimental pore pressure is dissipated during cohesive marine muds (Fig. 9E; Coussot et al., 2002; Debris Flows debris fl ow motion over periods of as much Jeong, 2010; Jeong et al., 2010) and due to Strongly coherent debris fl ows are prone to as several hours (Fig. 11). It is therefore likely positive feedbacks associated with mixing and hydroplaning at their head, and undergo rela- that the debris will still be partly or wholly dilution (see discussion of Potential Bifurca- tively little mixing with surrounding seawater. liquefi ed for signifi cant periods after it is depos- tion in Mixing and Flow Behavior). Deceler- Mixing occurs mainly at the head of the debris ited. It is therefore unsurprising that overly- ation may lead to increased viscosity, which in through erosion of single grains at lower speeds, ing ripple cross-laminated sands often founder turn causes further deceleration of the fl ow, and and chunks or a sheared layer of material at into underlying cohesive debrite (Butler and therefore even higher viscosity (Coussot et al., higher speeds. This material travels backward, McCaffrey , 2010). 2002). Increased velocity may reduce viscosity, and may partly settle back into the body. The thereby promoting even faster fl ow. This may fi ner grained material forms a dilute turbid- Turbulence Damping or Extinction favor late-stage transformation from turbulent ity current above the body. Strongly coherent to laminar fl ow as fl ows decelerate. debris fl ows tend to produce rather small vol- Small amounts of cohesive mud can dampen ume and dilute turbidity currents, which initially turbulence, especially at slower fl ow velocities Experimental Debris Flows of Higher trail behind the debris fl ow. However, the dilute (Fig. 8; Baas et al., 2009; Sumner et al., 2009). or Lower Coherency turbidity currents tend to run out beyond the Baas et al. (2009) reported that just 0.75% vol- location where the debris fl ow comes to a halt. ume kaolin (a rather weak clay mineral) was Experiments in which mixtures of mud and The highly coherent debris fl ow tends to extend suffi cient to dampen turbulence locally in a fl ow sand are released into fl ume tanks illustrate the back continuously to the point at which it ini- moving at <50 cm/s (Fig. 8); however, for fl ow dynamics of cohesive debris fl ows, and how tiates, although it can locally display tensional traveling at 1 m/s, turbulence was damped when debris fl ows generate turbidity currents through cracks (Mohrig et al., 1998). Sand is trapped kaolin volume concentrations exceeded ~6% mixing with surrounding water (Hampton, within the laminar plug composing the body of (Fig. 8). Many submarine fl ows carry signifi cant 1972; Mohrig et al., 1998; Marr et al., 2001; the debris fl ow, and does not segregate.

478 Geosphere, June 2013

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/3/460/3343593/460.pdf by guest on 28 September 2021 Hybrid submarine sediment ﬂ ows

Moderately Coherent Experimental This scaling can be expressed by the ratio An even clearer understanding of how debris Debris Flows of the sediment yield strength and the dynamic fl ow dynamics scale up from laboratory to full- There is a gradual continuum between strongly pressure at the fl ow front. However, other key scale submarine examples is a priority for future coherent and weakly coherent debris fl ows, with properties will not scale up in the same way as experimental studies. the degree of mixing with surrounding seawater this ratio. This can be illustrated by considering increasing progressively, so that progressively an experimental debris fl ow with yield strength GENERALIZED MODEL FOR more voluminous and denser turbidity cur- of 30 Pa and speed of 50 cm/s, and a submarine HYBRID FLOWS WITH INCREASING rents are produced. The head of the debris fl ow debris fl ow with yield strength of 1080 Pa SEDIMENT STRENGTH becomes turbulent as the relative magnitude of traveling at 3 m/s; both have similar ratios of shear stresses increase. yield strength to frontal dynamic pressure. The Field observations combined with the experi- much stronger submarine fl ow will be able to mental and theoretical analysis produce a gen- Weakly Coherent Experimental Debris Flows support much larger grains or clasts. The criti- eralized model for cohesive debris fl ows and The morphology, dynamics, and deposits of cal shear stress needed to erode the surface of hybrid fl ows (Fig. 12). The classifi cation is weakly coherent debris fl ows differ signifi cantly a sediment mixture (typically 0.5–2 Pa) can be based on increasing strength of the sediment from those of strongly coherent debris fl ows. orders of magnitude lower than its yield strength mixture that tends to produce more strongly The head and frontal part of the body become (Talling et al., 2002), and much higher surface coherent debris fl ows (Marr et al., 2001). The increasingly turbulent, and this turbulence pro- erosion rates may characterize the faster sub- yield strengths quoted for each type of debris motes much more effective mixing with sur- marine debris fl ow. Laboratory studies of mixing fl ow are approximate, as fl ow behavior will also rounding seawater. The trailing part of the body processes have tended to concentrate on pro- depend on other factors such as fl ow speed and develops a pronounced interface separating a cesses that occur near the head, rather than on the dynamic pressures exerted on the sediment basal high-density fl ow from an upper, much the upper surface of the debris fl ow body. This mixture. Decreasing shear will tend to decrease, more dilute, turbidity current (Marr et al., 2001; is partly because the head of a small-volume sometimes substantially, the viscosity and Mohrig and Marr, 2003; Ilstad et al., 2004; experimental debris fl ows forms a disproportion- strength of shear thinning muddy fl uids. Breien et al., 2010). The experiments of Ilstad ately large fraction of the debris fl ow, and mixing et al. (2004) and Breien et al. (2010) showed along the body may be more important for much Higher Strength Clast-Rich Debris Flows that excess pore pressures in the high-density larger submarine fl ows (Talling et al., 2002). layer are suffi cient to support the sediment. Differences in shear rates between laboratory Very high strength (>100 Pa) debris fl ows Breien et al. (2010) showed how the frontal part experiments and full-scale submarine fl ows may produce debrites that are often tens of meters of the high-density layer comprises a liquefi ed also be important because of the shear thinning thick (Fig. 12A) and that support large clasts layer from which sand settles out, incrementally rheology that characterizes many marine clays (Fig. 9). Such thick cohesive debrites tend to be depositing a layer of clean massive sand. Their (Jeong, 2010; Jeong et al., 2010). This shear restricted to the continental slope (e.g., Laberg experiments show how the rear of the dense thinning rheology can cause substantial (orders and Vorren, 2000). However, a few very thick layer comprises the same muddy sediment that of magnitude) decreases in the viscosity of the debris fl ows have run out onto low-gradient was introduced into the tank, as it is relatively debris fl ow at high shear rates, or increases (<~0.2°) distal fans and basin plains to produce protected from mixing. Sediment in the rear part in viscosity as shear rates decrease (Fig. 9E; megabeds (Labaume et al., 1987; Kleverlaan, of the dense fl ow layer is supported by muddy Coussot et al., 2002; Jeong, 2010; Jeong et al., 1987). High-strength (10–100 Pa) cohesive matrix strength rather than by liquefaction, and 2010). A decrease in viscosity can lead to even debris fl ow will generally produce similar but deposits muddy sand en masse when it comes faster fl ow, or an increase in viscosity can cause thinner (0.5–3 m) deposits on continental slopes to a halt (Breien et al., 2010). The frontal part further fl ow deceleration. These positive feed- (Tripsanas et al., 2008). of the dense layer that deposits clean sand is backs can therefore potentially lead to a bifur- Very high and high-strength debris fl ows are continuously being fed from the rear part of the cation in fl ow velocity and fl ow behavior that unlikely to become turbulent, even on steeper dense layer, such that the two parts of the layer will not be seen clearly in short-lived laboratory gradients (Fig. 9), and their deposits tend to are closely coupled. experiments. Other scaling issues include the extend back to the vicinity of the original slope relatively short distance that experimental fl ows failure. Mixing with the surrounding seawater Scaling Up of Laboratory Debris Flow travel that, together with differential speeds in is limited, and any dilute turbidity currents that Experiments different parts of the event, will determine how are produced by mixing are of limited volume. The laboratory-scale fl ows are relatively thin a fl ow event organizes during longer run out in The dilute turbidity current tends to lag behind (<50 cm) and slow moving (40–100 cm/s), such the ocean. A fi nal important point is that sub- the head of the debris fl ow, but can run ahead that the dynamic pressures on the experimental marine debris fl ows will initially accelerate and of the debris fl ow, after the debris fl ow comes fl ows may be much less than in faster moving eventually decelerate to a standstill, such that to halt. Cohesive debrite intervals are overlain submarine debris fl ows. Sediment strength and the dynamic pressures undergone by a parcel by the trailing dilute turbidity current deposits, viscosity tend to be more important in the rela- of sediment may vary substantially through which form a relatively small fraction of the tively slow experimental fl ows. For example, time. The coherency of a debris fl ow may vary overall deposit. In some cases, the dilute turbid- the slow speed of the experimental fl ows tends therefore substantially through its history, with ity current may bypass across steeper gradient to reduce the importance of shear wetting at stronger coherency characterizing the fi nal seafl oor, and deposit in other locations further their base. This means that sediment mixtures stages of the fl ow. down the fan. with weaker strength in the laboratory fl ows will The importance of these scaling issues has The frontal part of faster moving higher tend to reproduce the behavior of faster fl owing only been partially addressed in previous pub- strength debris fl ows is prone to hydroplaning, mixtures with much greater strength in full- lications, even those that consider geometric if it is suffi ciently thick to move (Fig. 10). How- scale submarine examples. scaling and Froude or fl ow Reynolds numbers. ever, water overridden at the head may often be

Geosphere, June 2013 479

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/3/460/3343593/460.pdf by guest on 28 September 2021 Talling

A. Very-strong or strong (100 to > 1,000 Pa) debris flow deposit extends back up slope to near initial slope failure relatively small & dilute turbidity current as limited mixing thin or absent capping turbidite 1 to 50 m debris flow laminar, viscous, and slow moving (unless hydroplaning)

Thick (10’s m) debris flows can run out across low gradients to form Mega-Beds, turbidity current can bypass whose long run is due to basal lubrication to deposit further down slope

B. Intermediate strength (100 to ~ 5 Pa) debris flow debris flow can frond-like shape bypass proximal areas with abrupt pinchout long run out mainly due to low yield strength

plug can travel longer way ~1 m with limited shear mixing ~1 m

] 1-2 m weakly turbulent laminar plug or laminar sediment charged mixture forerunning turbidity current increasing mixing produces name for initial flow depends on terminology adopted or late stage settling from plug extensive thin low density turbidite

C. Low strength (0.1 to ~5 Pa) sediment mixtures Debrite distribution cohesive debris flow protected mixing feeds high density depends on basin shape C-1 from mixing at rear of event turbidity current

greater mixing and turbidite forms greater fraction of bed debris flow forms via bypass of debris flow C-2 flow transformation or flow transformation debris flow forms locally & is slow to avoid mixing 1–2 m debrite absent high-density ] 1–2 m fully turbulent turbidite base laminar plug

D. Very low strength (< 0.1 Pa) fluid mud

flow forms locally and must ponded mud deposit be slow to avoid mixing in basinal lows

fully turbulent ] 1–2 m laminar plug

Figure 12. Generalized models for the behavior and deposits of submarine cohesive debris ﬂ ows. (A) Very strong or strong cohesive strength. (B) Intermediate cohesive strength. (C) Low cohesive strength. (D) Generalized model for an even lower cohesive strength ﬂ uid mud layer.

480 Geosphere, June 2013

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/3/460/3343593/460.pdf by guest on 28 September 2021 Hybrid submarine sediment ﬂ ows

mixed with the overlying debris fl ow, to form a turbidite. The basal turbidite clean sand may 2004). This much wider review of hybrid bed basal layer of low sediment concentration. This at times be relatively thick, and the debrite can occurrence in 3 modern systems, 14 ancient out- basal layer may lubricate the fl ow, allowing transition downbasin into muddy silt. The low- crop sequences, and more than 5 ancient subsur- increased speeds and run out. est strength clast-poor debris fl ows may gener- face sequences strongly supports both conclu- ate deposits that form bulls-eye patterns in basin sions. Almost all of the hybrid beds described Intermediate-Strength Clast-Rich lows, rather than having digitate (fi ngered) here are from distal lobe (fan fringe) settings Debris Flows planform shapes seen in intermediate-strength with relatively infrequent or shallow bed amal- debris fl ows (Fig. 12C). Any water overridden gamation, or basin-plain settings with almost Intermediate-strength (~5–100 Pa) debris at the turbulent fl ow front is rapidly assimilated no bed amalgamation. The only exceptions fl ows produce thinner (1–2 m) cohesive debrites into the fl ow, and these fl ows therefore do not are one of the two types of hybrid beds (fi ner in distal locations on submarine fans, beyond hydroplane. grained muddy sandstones) described by Syl- the steeper continental slope (Fig. 12B). The The debrite is absent in proximal deposits, vester and Lowe (2004) in Carpathian outcrops debrites are commonly clast rich, and clasts can and this may be a combination of two distinct inferred to represent a channel fi ll. Haughton be >1 m in length if they are of lower density processes that are not mutually exclusive. First, et al. (2009) also stated that hybrid beds occur than the matrix. The sediment-charged fl ow is their low strength ensures that these fl ows may in the upper part of channel-fi ll deposits in a laminar during the later depositional stages of often be initially fully turbulent on steeper core from the Paleocene Schiehallion oil fi eld, fl ow, but it may be initially weakly turbulent slopes, such that the debris fl ow forms via fl ow located west of the Shetland Islands, although on steeper gradients. Hydroplaning may occur transformation from turbidity current (Figs. 6 a detailed description of these hybrid beds was at the head, but the overridden water may be and 12C; Sumner et al., 2009). This transfor- not provided. thoroughly mixed into the body. Relatively long mation will begin in slower moving parts of Generally thicker (>3 m to tens of meters) run out distances can result only from the lower the fl ow, such as at the rear. The resulting weak cohesive debris fl ow deposits are relatively strength of the debris fl ow, without the need for cohesive debris fl ow may travel only for a short common on proximal continental slopes, but hydroplaning. In some cases, these intermedi- distance, at a slow speed, because such low- turbidites encasing these deposits are poorly ate-strength debris fl ows produce deposits with strength debris fl ow mixes easily with surround- developed or absent (Tripsanas et al., 2008). intricate digitate planform shapes (Fig. 12B). ing seawater (Fig. 12). Second, the low coher- This may be due to the higher strength of such Limited shear mixing may lead to the forma- ency of these fl ows may ensure that mixing is debris fl ows that reduces mixing with seawater tion of low-volume dilute turbidity currents, as especially vigorous at the front of the fl ow (Marr and turbidity current generation. However, it is was the case for higher strength debris fl ows, et al., 2001), generating a high-density turbidity more likely that turbidity currents that are gen- such that fi ne-grained turbidite composes a current (or a dense fl uidized sand layer) in front erated bypass the steeper continental slope and small part of the overall deposit. For example, of the trailing cohesive debris fl ow (Breien et al., deposit farther from the source. Examples of the Mississippi fan debrites pass abruptly into 2010). This, together with an overlying dilute megabeds with very thick cohesive debrite and very thin muddy turbidites. However, as mix- turbidity current, may protect the debris fl ow turbidite sand intervals are also found in basin ing increases progressively larger volume dilute composing the rear of the event from mixing. plain or lobe settings, rather than on continen- turbidity currents may be generated, and these The forerunning turbidity current may be gen- tal slopes. turbidity currents may run out well beyond the erated by this progressive mixing of the initial debris fl ow (as seen in Bed 2.5 of the Marnoso- debris fl ow, explaining why the cohesive debrite Consistent Level at Which Cohesive arenacea Formation; Talling et al., 2012b). is almost always found at the boundary between Debrites Occur Within a Hybrid Bed Debrites may be underlain by rather thin mas- high-density and low-density turbidite (Fig. 2). sive clean sand intervals, which tend to pinch A potentially important observation is the out at the same place as the overlying debrite. Very Low Strength Fluid Mud rela tively consistent level at which cohesive The basal sand may result from initial deposi- debrite intervals occur within hybrid beds tion from high-density turbidity current, gener- A further decrease in cohesive strength (Fig. 2). The cohesive debrite interval is almost ated by mixing at the head of the fl ow that is (<0.1 Pa) produces a transition into fl uid mud always underlain by massive, or occasionally undergoing relatively high dynamic pressures. layers (Fig. 12D), the strength of which is insuf- planar laminated, clean sandstone (broadly Alternatively, the thin basal clean sand may be fi cient to support sand. Low-strength debris equivalent to T or T intervals; see later dis- formed by late-stage settling of sand from the A B fl ows (that carry sand) may share many aspects cussion about its origin). The only examples debris fl ow plug, possibly at times even after the of fl uid mud behavior, such as ponding in of cohesive debrites underlain by ripple cross- debris fl ow has stopped moving (Sumner et al., basinal lows and late-stage transformation from laminated sand (T ) are the fi ner grained cohe- 2009), or due to shear shinning effects during C turbulent to laminar fl ow. The behavior of very sive debrites of Sylvester and Lowe (2004). the fl ow. low strength debris fl ows will also tend toward Clast-poor (lower strength) debrite intervals in that of turbidite mud deposition (McCave and Low-Strength Debris Flows the Marnoso-arenacea Formation are sometimes Jones, 1988; Talling et al., 2012a). directly underlain by dune-scale (20–90 cm Although their cohesive strength is strong DISCUSSION wavelength) cross-bedding. Such large-scale enough to carry sand, these lower strength (0.1 bedforms are often inferred to originate from to ~5 Pa) debris fl ows generally produce fi ner Depositional Setting of Hybrid Beds relatively dilute fl ow (Southard, 1991), but Baas grained muddy sandstones with few, if any, et al. (2011) showed how they may also origi- small clasts (Fig. 12C). Their deposits tend to Previous studies suggested that hybrid beds nate below more concentrated mud-rich fl ows. be relatively thin. Mixing with surrounding sea- are common in distal depositional settings, and Cohesive debrite intervals are not observed water is more effective, such that much of the that they are typically absent in more proximal within thinly (<30 cm) bedded turbidites char-

deposit comprises high-density and low-density settings (Haughton et al., 2003; Talling et al., acterized by TC, TD, or TE intervals in any of

Geosphere, June 2013 481

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/3/460/3343593/460.pdf by guest on 28 September 2021 Talling

these examples described here. However, cohe- of high-density turbidite and clast-rich muddy margins on steep (>1°) slopes (Pierson et al.,

sive debrites are commonly overlain by TC, TD, debrite sand can change rapidly, despite the 1990; Revellino et al., 2004). Bypass by sub-

or TE intervals deposited from relatively dilute overall bed thickness remaining near constant. aerial debris fl ows typically occurs for <1 or turbidity currents, from which they are typically This does not appear to result from local nests 2 km (Revellino et al., 2004), although subaerial separated by a grain size break. of clasts within muddy debrite sandstone, as debris fl ows have traveled as much as 55 km It therefore appears that cohesive debrites clast-free areas comprise clean turbidite sand. with limited deposition (Pierson et al., 1990). consistently mark the boundary between depo- A similar geometrical relationship occurs over Debris fl ow deposits in areas of subaerial bypass sition from high-density and low-density turbid- somewhat longer distances for hybrid beds with can contain boulders, and are typically tens of ity currents (Figs. 1 and 2). As a bed is built up clast-rich cleaner sand debrites in the Marnoso- centimeters thick (Pierson et al., 1990; Revellino progressively, the vertical bed structure records arenacea Formation (Talling et al., 2012c). This et al., 2004). To explain the location of cohesive the longitudinal fl ow structure at that site. The type of bed geometry could suggest that clast- debrites in submarine fl ow deposits, bypass of position of the debrite may be due to the speed rich debris fl ows tend to erode out a similar much larger sediment volumes would need to of the cohesive debris fl ow being consistently thickness of previously deposited high-density occur on much lower gradients. Such bypass intermediate between that of high-density and turbidite, but it is surprising that the thickness may be recorded by grain size breaks between low-density turbidity currents, across the same of eroded and deposited sediment is so similar. turbidite sand and mud in proximal deposits, but fl ow path. The debris fl ow therefore arrives at Alternatively, this geometry could indicate that there are typically no other signs of large-vol- the site after the high-density turbidity current, there is a close link between cohesive debris fl ow ume bypass in proximal turbidites. This might but before the arrival of low-density turbid- deposition and high-density turbidite deposition, be consistent with the generally larger volumes ity current. Such a relationship might also be such that the two processes deposit similar sedi- and lower yield strengths of submarine debris consistent with lubrication of the debris fl ow ment thicknesses in adjacent locations. fl ows, which are generally able to travel across by dewatering of previously deposited high- Correlation of hybrid beds containing clast- lower seafl oor gradients (<0.1°). density turbidite sand, and a lack of lubrication poor debrites shows that the debrite can some- This model would be consistent with a debrite by underlying low-density turbidite sand. How- times be on top of a relatively constant thickness matrix, which is often rich in organic matter, ever, this explanation is not favored because of high-density turbidite, such that the overall coming primarily from shallow water. However, some basal high-density turbidite intervals have bed thickness changes signifi cantly. This differ- the model needs to explain why the clasts within undeformed laminations, and because strongly ent (noncompensating) internal bed geometry that matrix typically lack abundant organic contorted laminations show that low-density suggests that the clast-poor debris fl ows are less material, and therefore often have a composition

(TC) turbidite sand can also undergo pervasive erosional, or that deposition from the clast-poor different from the organic-rich matrix. A poten- syndepositional dewatering. debrite and high-density turbidity current are tial explanation is that the original slope failure The debris fl ows only occur in fl ow events not closely linked. was of weaker strata rich in organics and inter- that also contain high-density turbidity currents, bedded more resistant intervals. Such interbed- as well as trailing low-density turbidity currents. Origins of Cohesive Debris Flow in ded sequences of weaker organic-rich turbidites This relationship could be due mainly to the Hybrid low Events and stronger hemipelagic mud can occur on longer run out of low-density turbidity currents Four general types of model can be proposed continental slopes offshore major river deltas, than either debris fl ow or high-density turbidity for the origin of cohesive debris fl ows within due to cyclic changes in sea level that result in current. Confi nement of cohesive debris fl ows hybrid fl ow events. episodic deposition of turbidite packages (e.g., in channels will also explain their paucity in Ducassou et al., 2008, 2013). levee sequences, which are dominated by thin Debris Flow Originates from Initial low-density turbidites. However, bed correla- Slope Failure (Model 1) Debris Flow Formed by Flow Transformation tion in the Marnoso-arenacea Formation shows The debris fl ow can originate from the same Due to Mud Eroded along Flow Path that clast-poor cohesive debrite can sometimes initial source as the rest of the hybrid fl ow, for (Model 2) replace thick intervals of ripple cross-laminated example from an initial slope failure. The debris A second hypothesis is that the cohesive low-density turbidite over a few kilometers in a fl ow component of the fl ow needs to bypass debris fl ows originated through local erosion of downfl ow direction (Talling et al., 2012b). This through more proximal parts of the fan, where muddy seafl oor sediment along the fl ow path. transition suggests that the presence of cohesive only turbidite sand is deposited. Bypass could The eroded muddy material causes turbulence debris fl ow can also sometimes suppress ripple occur if the matrix strength, and other support to be damped and local transformation from development in the low-density tail of the fl ow mechanisms such as excess pore pressure, is suf- turbidity current to debris fl ow. In this scenario, event (cf. Sylvester and Lowe, 2004). fi cient to keep the sand suspended. Unlike tur- the debris fl ow contains additional sediment that bidity currents, debris fl ows deposit en masse, comes locally from the seafl oor. Experiments Implications of Lateral Changes in Debrite and this could mean that they are more likely have shown how small increases in cohesive and Hybrid Bed Thickness to bypass sediment initially before deposition mud content can lead to the support of sand fi nally occurs. Larger clasts may segregate from grains (Sumner et al., 2009), and how increased Some fi eld observations suggest that deposi- a bypassing debris fl ow, potentially leaving a mud content can dampen fl ow turbulence (Baas tion by cohesive debris fl ows and high-density coarse lag of clasts in proximal areas of bypass et al., 2009). It is possible that a small amount of turbidity currents can on occasions be closely (Talling et al., 2007a). Hydroplaning could aid erosion could increase the mud content above a linked. Hybrid beds containing cohesive debrites bypassing, although this process will only affect threshold value, leading to fl ow transformation. often have thickness distributions very similar the very front of the fl ow, and turbulence can lead Clasts in the debris fl ow should come from to those of beds comprising only turbidite sand- to mixing of overridden water back into the fl ow. deeper water (see Ito, 2008). However, other stone in the same sequence. Hodgson (2009, fi g. Subaerial debris fl ows can leave behind rela- debrites contain exotic clasts that could not 3C therein) illustrated how the relative thickness tively thin coarse-grained deposits near their come from local erosion of the seafl oor (Talling

482 Geosphere, June 2013

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/3/460/3343593/460.pdf by guest on 28 September 2021 Hybrid submarine sediment ﬂ ows

et al., 2007b; Haughton et al., 2009). Erosion the theoretical analysis presented here shows to occur only in relatively weak debris fl ows, of muddy seafl oor sediment should generally how somewhat stronger debris fl ows could run and will form basal sand that terminates where decrease the relative abundance of organic out for long distances on low gradients, if not the debris fl ow stops. carbonaceous material, as seafl oor sediment diluted by shear mixing. The observation that contains less organic material in deeper water cohesive debrites consistently underlie rippled Reasons for Mobility of Thin Debris Flows locations. However, in many cases, the debrite intervals suggests that the debris fl ows were on Low Slopes intervals have more abundant organic mate- faster moving than the dilute turbidity currents Rather thin cohesive debrites within hybrid rial than adjacent turbidite mud and sand or that deposited the rippled interval. This in turn beds, and debrites lacking surrounding turbidite hemipelagic mud. Cohesive debrite is typically suggests that the debris fl ows were relatively (Migeon et al., 2010), can be found as much as underlain by a continuous layer of clean sand, fast moving, and could be formed from initially several hundred kilometers from their source which forms a barrier between the debrite and fast moving and turbulent fl ow. in areas of very low gradient (0.05°) seafl oor areas of eroded seafl oor. Unless this clean sand (Talling et al., 2004). The mobility of thin sub- forms through late-stage settling of sand from Debris Flows Triggered by Local Margin marine debris fl ows therefore tends to be much the debris fl ow, it is not clear as yet how such a Failure (Model 4) greater than most of their subaerial counterparts. continuous barrier would form if the debris fl ow It is unlikely that the cohesive debris fl ows This contribution has shown how thin and was generated by seafl oor erosion. were generated by loading and failure of basin low-strength submarine debris fl ows can sup- margin slopes by an initial turbidity current port sand, but still run out across very low- Debris Flows Formed by Flow (McCaffrey and Kneller, 2001). This process gradient seafl oor (Fig. 9B). Therefore, debris Transformation from Turbidity Current could not easily explain the organic-rich matrix fl ows need not necessarily be lubricated at their Without Erosion (Model 3) of many cohesive debrites, or the exotic mud base to achieve these run out distances. In some A third model is that the cohesive debris fl ow clasts seen in certain debrites (Talling et al., cases, submarine debris fl ows may carry very results from fl ow transformation from an ini- 2004; Haughton et al., 2009), and in some large clasts that are positively buoyant to the tially turbulent part of the fl ow (turbidity current, cases it can be shown that the debrites are not fringes of submarine fans (Talling et al., 2010; according to our defi nitions), but without the located near basin margins or topographic highs Fig. 5B). The conditions for hydroplaning may need for local erosion of muddy seafl oor. In this (Talling et al., 2007b). occur commonly (Fig. 10), but water overridden model, fl ow deceleration and reduced shearing by such hydroplaning fl ows may often be mixed of the muddy fl uid lead to fl ow transformation Dependence on Terminology for back into the body of the debris fl ow. as the existing mud within the fl ow increases Models 1 and 3 Haughton et al. (2003, 2009) proposed that fl ow viscosity and dampens turbulence at slower The difference between two of these models dewatering of the clean basal sand often plays fl ow speeds. Experiments suggest that such fl ow is really a matter of the terminology used to an important role in the long run out of overlying transformation may be a general characteristic describe an initial sediment laden dense fl ow. debris fl ow. In some of the examples described of muddy fl ows, the amount of mud control- If an initial dense turbulent suspension is called here this process did not occur, as cross-lamina- ling the speed at which transformation occurs. a turbidity current, then the debris fl ow forms tion or planar lamination in the underlying clean If transformation occurs before all the sand has via fl ow transformation in model 3. If the initial sand is not deformed (Fig. 1; Talling et al., 2004, settled out of the fl ow, then muddy debrite sand sediment-charged but turbulent initial suspen- 2007b). Haughton et al. (2003, 2009) described will result. sion was called a debris fl ow, then the debris a series of hybrid beds in which the underlying Debrites containing large clasts are less likely fl ow forms from the initial failure (model 1). clean sand is injected and preserves evidence of to have been deposited in this way, as the large dewatering. It is diffi cult to determine whether clasts would need to be carried within the ini- Origins of Basal Clean Sand this evidence means that dewatering played a tially turbulent fl ow. This might only occur if major role in debris fl ow motion, or whether turbulence was very strong (in which case one Dune cross-bedding, strong normal grading, the beds result from rapid loading, because it is might expect the clasts to rapidly break up), or if and planar laminations provide clear evidence unknown whether such dewatered clean sands the turbulent sediment mixture was suffi ciently of deposition from a forerunning turbidity cur- terminate at the same location as the debrite. dense that clasts were nearly buoyant. This rent in some cases (Talling et al., 2007b). How- However, basal dewatering of the clean sand process is more likely to produce clast-poor ever, it is less easy to determine unambiguously is apparent in the Haughton et al. (2003, 2009) debrites. This transformation process might whether massive clean sandstone is deposited examples. occur fi rst in slower moving parts of the fl ow, incrementally from a laminar dense fl uidized ensuring that the transformation may tend to fl ow of the type produced in the experiments of Implications for Petroleum Reservoirs start at the back of the fl ow. This could explain Breien et al. (2010), or from a turbulent high- why clast-poor (low strength) debrites are often density turbidity current. Clean sand layers Mud-rich cohesive debrite sandstone inter- underlain by thick turbidite sand, deposited formed from dense fluidized layers might vals have much lower permeability than clean from the still turbulent front of the fl ow event. pinch out more abruptly once the debris fl ow sand deposited by turbidity currents (Amy et al., Two variants of this model can be proposed. stops, while the run out of high-density turbid- 2009). This means that the debrite sandstone First, the cohesive debris fl ow may run out for ity current may be less connected to debris fl ow intervals have unfavorable reservoir quality and long distances after formation. Second, the motion. baffl es to subsurface fl ow, and it is important debris fl ow tends to form close to the site of Relatively thin massive graded or ungraded to predict their geometry to extract oil and gas debrite deposition. Very weak debris fl ows will clean sand layers could potentially be formed effectively. Hybrid beds are common in sys- tend to mix effi ciently with surrounding sea- by sand settling from the debris fl ow plug at a tems worldwide, and most beds in some subsur- water, even at low speeds, favoring formation late stage, or after the debris fl ow has stopped face cores (e.g., Haughton et al., 2009) can be locally near the site of deposition. However, (Sumner et al., 2009). Such a process will tend hybrid beds.

Geosphere, June 2013 483

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/3/460/3343593/460.pdf by guest on 28 September 2021 Talling

Controls on Hybrid Bed Location, Shape, son, 2009; Pyles and Jennette, 2009). Channel- an excellent starting point for describing the and Extent lobe progradation may lead to intervals rich in basic structure of hybrid beds. Figure 2A pro- There is a strong tendency for hybrid beds to hybrid beds, overlain by more proximal lobes of vides a summary of hybrid bed character seen in occur in the intermediate parts of larger systems, channel-levee facies that lack hybrid beds, and this wider collection of examples, and includes as cohesive debrites are absent in more proxi- underlain by very thin bedded turbidites from variations in hybrid bed character and examples mal and the most distal locations. The more the most distal fan locations. Retrogression of of downfl ow bed geometries (Fig. 2B). I empha- proximal locations comprise only thick-bedded the channel lobe may lead to the inverse succes- size here the variability in hybrid bed types as clean sandstones formed by high-density turbid- sion, although this may be rarer due to abrupt cohesive debris fl ow strength increases, and how ity currents, while the distal areas are typically lobe abandonment because of channel avulsion. fl ow evolution leads to different bed geometries. dominated thin turbidites deposited by dilute Larger scale progradation of the entire fan sys- turbidity currents. tem may also control hybrid bed frequency over Basal Turbidite Sand Interval (H1) Cohesive debris fl ow deposition is more longer time scales. Hybrid beds can show very Haughton et al. (2009) ascribed the basal strongly affected by small changes in seafl oor little if any clustering in situations where the clean sand interval (H1; Fig. 2) to deposition gradient than turbidity current deposition (Fig. depositional setting did not fl uctuate strongly, by a high-density turbidity current. This is the 3B; Talling et al., 2007a; Wynn et al., 2010). such as the thick basin plain sequence of the depositional mechanism in most cases (see dis- This means that the location of debrite intervals (inner part of the) Marnoso-arenacea Formation cussion of Origins of Basal Clean Sand), but is controlled by (often subtle) changes in basin (Talling et al., 2012b). basal clean sand may be formed in a smaller morphology. For example, basin morphology Hybrid beds are associated with larger vol- number of cases by late-stage settling of sand can determine whether debrites form the fringes ume fl ows that include a high-density turbidity from the debris fl ow plug (Sumner et al., 2009), or core of a sequence. Debrites on the Nile fan current component, and tend to be lacking in or incrementally by a forerunning laminar form a fringe around the end of the upstanding small dilute turbidity currents. This means that dense liquefi ed fl ow of the type observed in the ridge formed by the channel system (Migeon clustering of thick high-density turbidites will experiments of Breien et al. (2010). Haughton et al., 2010); this is because fl ows that exited tend to be associated with clustering of hybrid et al. (2009) emphasized dewatering of the the channel underwent a radial break in slope, beds within those thicker bedded packages; this dense basal sand as a mechanism for long run and therefore formed debrite in a fringe. In the can result from either channel-lobe migration or out. Here I show that some basal sand intervals Agadir Basin, cohesive debrite is found with a fl uctuations in sea level. have not dewatered and contain undeformed bulls-eye pattern in the basin’s two fl attest areas Haughton et al. (2009) suggested that cluster- planar or dune-scale lamination, and that low- (Fig. 3). The Agadir Basin lacks an upstanding of hybrid beds may result commonly from strength cohesive debris fl ows can potentially ing channel-levee ridge, and debrite therefore periods in which systems are out of grade. Such a run out for very long distances without the need occurs beyond subtle (0.05°–0.02°) breaks in model infers that hybrid beds are formed primar- for basal lubrication. slope at the basin center. ily by erosion of the muddy seafl oor along parts The location and shape of cohesive debrites of the fl ow path (model 2) that are periodically Banded Interval (H2) will also depend on the relative strength of a out of grade. The fi eld observations reviewed An interval with repeated banding separated debris fl ow (Fig. 12). Very high strength debris here suggest that a signifi cant number of cohe- the basal high-density turbidite from the over- fl ows tend to produce thick clast-rich deposits sive debris fl ows, rich in organic material that lying cohesive debrite in the Haughton et al. in a single lobe, typically on the continental most likely came from initial slope failure, are (2009) model. This transition can be abrupt and slope (Laberg and Vorren, 2000). Intermedi- not formed by local erosion along the fl ow path. occur gradationally over a few centimeters, or ate-strength debris fl ows can produce thinner Laboratory experiments show that any fl ow that more rarely occur in a series of steps with pro- clast-rich debris fl ows with a digitate frond-like contains a signifi cant amount of cohesive mud gressively increasing mud content. Dune-scale planform shape, which may lack thick encasing content can undergo late-stage transformation cross-bedding is also rarely observed at this turbidite sand, such as on the Mississippi and from turbulent to laminar fl ow. This cohesive boundary (Figs. 1 and 2). Repeated banding of Nile fans (Fig. 4). The digitate planform shape mud fraction can be present from the start of the the type depicted in the Haughton et al. (2009) is diagnostic of debris fl ow deposition, and can event, or be picked up en route via erosion. It is idealized model is rare, although seen in other be visible in attribute maps of high-resolution not clear whether “out of grade” refers to sub- locations (Sylvester and Lowe, 2004). As yet, it three-dimensional seismic refl ection data. As marine canyons cut into the continental slope, has only been shown to form very thick inter- debris fl ow strength decreases, debris fl ow or local topographic anomalies further down the vals in the Britannia Formation in the North Sea deposits will tend to run out further from source, fan. Submarine canyons tend to be areas of net (Fig. 1; Lowe and Guy, 2000). The Haughton although the run out is also dependent on factors erosion over long periods, and might be said to et al. (2009) model is therefore amended to such as debris fl ow volume. Very low strength always be out of grade. The available evidence show the variability of the transitional H2 inter- (clast poor) debris fl ows may lack a digitate therefore does not provide strong support for a val, and to show that repeated banding is not the planform shape and may pond in a bulls-eye model in which local erosion on the fan deter- norm (Fig. 2). shape in the fl attest parts of a basin (similar to mines clustering of hybrid beds, although such a fl uid mud) (Figs. 3 and 12). process could potentially occur in some locations. Cohesive Debrite Interval (H3) The cohesive debrite interval contains clasts Stratigraphic Clustering of Hybrid Beds Comparison to the Haughton et al. (2009) of variable types, or no clasts, as described by Cohesive debrites tend to occur in the distal Classifi cation of Hybrid Beds Haughton et al. (2009). Sand injection does not (but not most distal parts) of submarine fans. occur in many examples reviewed here, although Patterns of fan progradation may therefore be Haughton et al. (2009, fi g. 3 therein) provided it is common in the hybrid beds described by an important control on the overall distribution a graphic log showing an idealized fi ve-part Haughton et al. (2009) in the North Sea (see dis- of hybrid beds in a stratigraphic interval (Hodg- hybrid event bed (H1–H5); their model provides cussion of Hybrid Beds in Subsurface Cores).

484 Geosphere, June 2013

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/3/460/3343593/460.pdf by guest on 28 September 2021 Hybrid submarine sediment ﬂ ows

The Haughton et al. (2009) model is extended fl ow transformation near the fi nal site of debrite Amy, L.A., Talling, P.J., Peakall, J., Wynn, R.B., and Arzola Thynne, R.G., 2005, Bed geometry used to test rec- herein by showing how increasing cohesive deposition, and emplacement must be gentle if ognition criteria of turbidites and (sandy) debrites: debrite strength infl uences that character of the they are not to mix with surrounding seawater. Sedimentary Geology, v. 179, p. 163–174, doi:10.1016 hybrid bed (Fig. 12). As cohesive strength is further reduced there is /j.sedgeo.2005.04.007. Amy, L., Talling, P.J., Edmonds, V.O., Sumner, E.J., and a transition into fl uid mud layers that lack sand, Leseuer, A., 2006, An experimental investigation of Upper Laminated Low-Density Turbidite (H4) and the processes that deposit turbidite mud. sand-mud suspension settling behaviour: Implications and Turbidite Mud (H5) The location and shape of cohesive debrites are for bimodal mud contents of submarine fl ow deposits: Sedimentology, v. 53, p. 1411–1434, doi:10.1111 As noted by Haughton et al. (2009), the controlled strongly by subtle changes in seafl oor /j.1365-3091.2006.00815.x. uppermost part of the hybrid bed comprises fi ne- gradient. Cohesive debrites in hybrid beds may Amy, L.A., Peachey, S.A., Gardiner, A.A., and Talling, P.J., 2009, Prediction of hydrocarbon recovery from turbidite grained low-density turbidity current deposits form a fringe around upstanding channel-levee sandstones with linked-debrite facies: Numerical fl ow- (Fig. 2). In some cases the debrite is directly ridges, or occur in the central lowest part of basin simulation studies: Marine and Petroleum Geology, v. 26, overlain by turbidite mud. In other cases, the plains lacking such uplifted ridges. This allows p. 2032–2043, doi:10.1016/j.marpetgeo.2009.02.017. Arnott, R.W.C., and Hand, B.M., 1989, Bedforms, primary debrite is overlain by ripple cross-laminated cohesive debrite geometry to be predicted for structures and grain fabric in the presence of suspended sand, which have sometimes partly foundered subsurface oil and gas reservoirs, if basin fl oor sediment rain: Journal of Sedimentary Petrology, v. 59, into the underlying debrite. Planar laminated morphology is well understood. Small frac- p. 1062–1069. Baas, J.H., and Best, J.L., 2002, Turbulence modulation in intervals tend to overlie ripple cross-laminated tional changes in mud content lead to changes clay-rich sediment-laden fl ows and some implications for sand in the caps of hybrid beds reviewed here, in cohesive strength, viscosity, permeability, and sediment deposition: Journal of Sedimentary Research, v. 72, p. 336–340, doi:10.1306/120601720336. rather than occurring below ripple cross-lami- rates of excess pore pressure dissipation that Baas, J.H., and Best, J.L., 2008, The dynamics of turbulent, nated intervals (as in the Haughton et al., 2009, span several orders of magnitude. Debris fl ows transitional and laminar clay-laden fl ow over a fi xed model). The planar laminated intervals in the H4 may be strongly shear thinning; this, together current ripple: Sedimentology, v. 55, p. 635–666, doi: 10.1111/j.1365-3091.2007.00916.x. interval therefore correspond to the TD division with mixing processes, can lead to bifurcation Baas, J.H., Best, J.L., Peakall, J., and Wang, M., 2009, A rather than the TB division (Fig. 2). in fl ow behavior. Small amounts of mud can phase diagram for turbulent transitional and lami- dampen turbulence effectively , especially as nar clay suspension flows: Journal of Sedimentary Research, v. 79, p. 162–183, doi:10.2110/jsr.2009.025. CONCLUSIONS fl ow decelerates, and fl ow transformation may Baas, J.H., Best, J.L., and Peakall, J., 2011, Depositional be common. This ensures that hybrid fl ows and processes, bedform development and hybrid fl ows in rapidly decelerated cohesive (mud-sand) sedi- Hybrid beds are a major departure from pre- cohesive debris fl ows display a wide range of ment fl ows: Sedimentology, v. 58, p. 1953–1987, vious widely cited models for submarine fl ow fl ow behavior and deposit geometries. doi:10.1111/j.1365-3091.2011.01247.x. deposits that capture our understanding of the Bandini, P., and Sathiskumar, S., 2009, Effects of silt content ACKNOWLEDGMENTS and void ratio on the saturated hydraulic conductivity fl ows (e.g., Bouma, 1962; Lowe, 1982; Mutti, and compressibility of sand-silt mixtures: Journal of 1992; Mulder and Alexander, 2001). Cohesive This synthesis includes insights from a wide range Geotechnical and Geoenvironmental Engineering, debris fl ows are best classifi ed in terms of a con- of previous work, for which I am grateful. Lawrence v. 135, p. 1976–1980, doi:10.1061/(ASCE)GT.1943 -5606.0000177. tinuum of decreasing cohesive strength (Talling Amy (Tullow Oil) and Esther Sumner (University of Southampton) played major roles in complet- Bannerjee, I., 1977, Experimental study on the effect of deceleration on the vertical sequence of sedimentary et al., 2012a; Fig. 12). High-strength (~100 to ing laboratory experiments and fi eld work in the >1000 Pa) debris fl ows tend to produce clast- structures in silty sediments: Journal of Sedimentary Marnoso-arenacea Formation, the latter in collabo- Petrology, v. 47, p. 771–783. rich debrites that are relatively thick and extend ration with Giuseppe Malgesini (National Oceanog- Barker, S.P., Haughton, P.D.W., McCaffrey, W.D., Archer, back to near the site of original slope failure. raphy Centre, NOC) and Fabrizio Felletti (University S.G., and Hakes, B., 2008, Development of rheological They are mostly restricted to steeper continental of Milan). Field data collection in the Moroccan tur- heterogeneity in clay-rich high-density turbidity cur- bidite system benefi tted from the considerable efforts rents: Aptian Britannia Sandstone Member, UK Con- slopes, but sometimes form megabeds on basin of Russell Wynn, Michael Frenz, Doug Masson, tinental Shelf: Journal of Sedimentary Research, v. 78, plains, and tend to lack well-developed encasing James Hunt, Christopher Stevenson (all at NOC), p. 45–68, doi:10.2110/jsr.2008.014. Blackbourn, G.A., and Thomson, M.E., 2000, Britannia turbidite sand. Intermediate cohesive strength and others over many years. Sebastian Migeon, Jean Field, UK North Sea: Petrographic constraints on (~5 to ~100 Pa) debris fl ows often contain Mascle, and Emmanuelle Ducassou (Geoscience Lower Cretaceous provenance, facies, and the origin Azur) provided recently published information on the clasts, and may commonly reach the distal low- of slurry-fl ow deposits: Petroleum Geoscience, v. 6, Nile system. David Twichell (U.S. Geological Survey) p. 329–343, doi:10.1144/petgeo.6.4.329. gradient parts of submarine fans, to produce helped greatly in the analysis of Mississippi fan cores. Bouma, A.H., 1962, Sedimentology of some fl ysch deposits: debrites that are less than a couple of meters This work was part of a UK-TAPS (U.K. Turbidite A graphic approach to facies interpretation: Amster- thick. Clast composition shows that debris fl ow Architecture and Process Studies) project, funded dam, Elsevier, 168 p. by NERC (Natural Environment Research Council) Bowen, A.J., Normark, W.R., and Piper, D.J.W., 1984, can be very far-traveled, and meter-sized clasts Modelling of turbidity currents on Navy Submarine Ocean Margins LINK grants NER/T/S/2000/0106 can be rafted for long distances on low gradi- Fan, California Continental Borderland: Sedimentol- and NER/M/S/2000/00264, and cosponsored by ogy, v. 31, p. 169–185, doi:10.1111/j.1365-3091.1984. ents if the clasts are less dense than surrounding ConocoPhillips, BHP Billiton, and Shell. The reviews tb01957.x. fl ow. Low-strength (0.1–5 Pa) cohesive debris of Joris Eggenhuisen, David Piper, and an anonymous Breien, H., De Blasio, F.V., Elverhøi, A., Nystruen, J.P., and fl ows generally lack larger (greater than milli- reviewer were very much appreciated. Harbitz, C.B., 2010, Transport mechanisms of sand in deep-marine environments—Insights based on labora- meter) clasts. Intermediate- and low-strength REFERENCES CITED tory experiments: Journal of Sedimentary Research, debris fl ows may evolve from initially turbulent v. 80, p. 975–990, doi:10.2110/jsr.2010.079. and sediment-charged fl ows on steeper slopes Allen, J.R.L., 1985, Principals of physical sedimentology: Butler, R.W.H., and McCaffrey, W.D., 2010, The structural London, George Allen and Unwin, 272 p. geology of sediment entrainment in submarine mass in proximal areas, and their deposits are com- Amos, C.L., Sutherland, T.F., and Zevenhuizen, J., 1996, transport complexes: Outcrop perspectives: Geologi- monly restricted to hybrid beds in the distal parts The stability of fi ne-grained sediment in a subarctic cal Society of London Journal, v. 167, p. 617–631, doi:10.1144/0016-76492009-041. of submarine fans. Intermediate-strength clast- estuary: Sedimentology, v. 43, p. 1–19, doi:10.1111 /j.1365-3091.1996.tb01455.x. Cantero, M.I., Cantelli, A., Pirmez, C., Balanchandar, S., rich debris fl ows may run out for long distances Amy, L., and Talling, P.J., 2006, Anatomy of turbidite and Mohrig, D., Hickson, T.A., Yeh, T., Naruse, H., and on low gradients (as shown by clast composi- debrite sandstones based on long distance (120 × Parker, G., 2011, Emplacement of massive turbidites 35 km) bed correlation, Marnoso Arenacea Formation, linked to extinction of turbulence in turbidity cur- tions) without hydroplaning. Very low strength Northern Apennines, Italy: Sedimentology, v. 53, rents: Nature Geoscience, v. 5, p. 42–45, doi:10.1038 cohesive debris fl ows most likely form via local p. 161–212, doi:10.1111/j.1365-3091.2005.00756.x. /ngeo1320.

Geosphere, June 2013 485

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/3/460/3343593/460.pdf by guest on 28 September 2021 Talling

Carter, L., Milliman, J., Talling, P.J., Gavey, R., and Wynn, Harbitz, C.B., Parker, G., Elverhoi, A., Marr, J.G., Mohrig, Jackson, C.A.L., Adli Zakaria, A., Johnson, H., Tongkul, F., R.B., 2012, Near-synchronous and delayed initiation of D., and Harff, P.A., 2003, Hydroplaning of subaqueous and Crevello, P.D., 2009, Sedimentology, stratigraphic long run-out submarine sediment fl ows from a record debris fl ows and glide blocks: Analytical solutions and occurrence and origin of linked debrites in the West breaking river-fl ood, offshore Taiwan: Geophysical discussion: Journal of Geophysical Research, v. 108, Crocker Fm (Oligo-Miocene), Sabah, NW Borneo: Research Letters, v. 39, doi:10.1029/2012GL051172. 2349, doi:10.1029/2001JB001454. Marine and Petroleum Geology, v. 26, p. 1957–1973, Coussot, P., 1995, Structural similarity and transition from Haughton, P.D.W., 2010, Hybrid event beds in outer fan suc- doi:10.1016/j.marpetgeo.2009.02.019. Newtonian to Non-Newtonian behaviour for water-clay cessions—The ‘greywacke problem’ revisited: British Jeong, S.W., 2010, Grain size dependent rheology on the suspensions: Physical Review Letters, v. 74, p. 3971– Sedimentology Research Group Annual Meeting 2010 mobility of debris fl ows: Geosciences Journal, v. 14, 3974, doi:10.1103/PhysRevLett.74.3971. Abstract Volume, p. 21. p. 359–369, doi:10.1007/s12303-010-0036-y. Coussot, P., Nguyen, Q.D., Huynh, H.T., and Bonn, D., 2002, Haughton, P.D.W., Barker, S.P., and McCaffrey, W., 2003, Jeong, S.W., Locat, J., Serge, L., and Malet, J.P., 2010, Avalanche behaviour in yield stress fl uids: Physical ‘Linked’ debrites in sand-rich turbidite systems—Ori- Rheological properties of fi ne-grained sediment: The Review Letters, v. 88, 175501, doi:10.1103/PhysRevLett gin and signifi cance: Sedimentology, v. 50, p. 459–482, roles of texture and mineralogy: Canadian Geotechni- .88.175501. doi:10.1046/j.1365-3091.2003.00560.x. cal Journal, v. 47, p. 1085–1100, doi:10.1139/T10-012. De Blasio, F.V., Engvik, L., Harbitz, C.B., and Elverhøi, Haughton, P.D.W., Davis, C., McCaffrey, W., and Barker, Johns, D.R., Mutti, E., Rosell, J., and Seguret, M., 1981, A., 2004, Hydroplaning and submarine debris fl ows: S.P., 2009, Hybrid sediment gravity fl ow deposits— Origin of a thick redeposited carbonate bed in Eocene Journal of Geophysical Research, v. 109, CO1002, Classifi cation, origin and signifi cance: Marine and turbidites of the Hecho Group, south-central Pyrenees, doi:10.1029/2002JC001714. Petroleum Geology, v. 26, p. 1900–1918, doi:10.1016 Spain: Geology, v. 9, p. 161–164, doi:10.1130/0091 Ducassou, E., Mulder, T., Migeon, S., Gonthier, E., Murat, /j.marpetgeo.2009.02.012. -7613(1981)9<161:OOATRC>2.0.CO;2. A., Revel, M., Capotardi, L., Bernasconi, S.M., Mascle , Haughton, P.D.W., Davis, C., McCaffrey, W., and Barker, Johnson, A.M., 1970, Physical processes in geology: A method J., and Zaragosi, S., 2008, Nile fl oods recorded in deep S.P., 2010, Hybrid sediment gravity fl ows—Classifi - for interpretation of natural phenomena; intrusions in Mediterranean sediments: Quaternary Research, v. 70, cation, origin and significance: Reply: Marine and igneous rocks, fractures, and folds, fl ow of debris and p. 382–391, doi:10.1016/j.yqres.2008.02.011. Petroleum Geology, v. 27, p. 2066–2069, doi:10.1016 ice: San Francisco, California, Freeman, Cooper, 577 p. Ducassou, E., Migeon, S., Mulder, T., Murat, A., Capotondi, /j.marpetgeo.2010.06.012. Khripounoff, A., Vangriesheim, A., Babonneau, N., L., Bernasconi, S.M., and Mascle, J., 2009, Evolu- Heezen, B.C., and Ewing, B.C., 1952, Turbidity currents and Crassous, P., Dennielou, B., and Savoye, B., 2003, tion of the Nile deep-sea turbidite system during the submarine slumps, and the 1929 Grand Banks earth- Direct observation of intense turbidity current activity late Quaternary: Infl uence of climate change on fan quake: American Journal of Science, v. 250, p. 849– in the Zaire submarine valley at 4000 m water depth: sedimentation: Sedimentology, v. 56, p. 2061–2090, 873, doi:10.2475/ajs.250.12.849. Marine Geology, v. 194, p. 151–158, doi:10.1016 doi:10.1111/j.1365-3091.2009.01070.x. Heezen, B.C., and Ewing, M., 1955, Orléansville earthquake /S0025-3227(02)00677-1. Ducassou, E., Migeon, S., Capotandi, L., and Mascle, J., and turbidity currents: American Association of Petro- Khripounoff, A., Vangriesheim, A., Crassous, P., and Etou- 2013, Run-out distance and erosion of debris-fl ows leum Geologists Bulletin, v. 39, p. 2505–2514. bleau, J., 2009, High frequency of sediment gravity in the Nile deep-sea fan system: Evidence from litho- Higgs, R., 2010, Hybrid sediment gravity fl ows—Classifi - fl ow events in the Var submarine canyon (Mediterra- facies and micropalaeontological analyses: Marine & cation, origin and signifi cance: Comment: Marine and nean Sea): Marine Geology, v. 263, p. 1–6, doi:10.1016 Petroleum Geology, v. 39, p. 102–123, doi:10.1016 Petroleum Geology, v. 27, p. 2062–2065, doi:10.1016 /j.margeo.2009.03.014. /j.marpetgeo.2012.09.006. /j.marpetgeo.2010.06.013. Kilfeather, A.A., O’Cofaigh, C., Dowdeswell, J.A., van der Elverhøi, A., Norem, H., Andersen, E.S., Dowdeswell, J.A., Hiscott, R.N., 1994, Traction-carpet stratifi cation in turbi- Meer, J.J.M., and Evans, D.J.A., 2010, Micromor- Fossen, I., Hafl idason, H., Kenyon, N.H., Laberg, J.S., dites; fact or fi ction?: Journal of Sedimentary Research, phological characteristics of glacimarine sediments: King, E.L., Sejrup, H.P., Solheim, A., and Vorren, T., v. 64, p. 204–208, doi:10.1306/D4267D57-2B26-11D7 Implications for distinguishing genetic processes of 2007, On the origin and fl ow behavior of submarine -8648000102C1865D. massive diamicts: Geo-Marine Letters, v. 30, p. 77–97, slides on deep-sea fans along the Norwegian–Barents Hiscott, R.N., and Middleton, G.V., 1979, Depositional doi:10.1007/s00367-009-0160-8. Sea continental margin: Geo-Marine Letters, v. 17, mechanics of the thick-bedded sandstones at the base King, E.L., Hafl idason, H., Sejrup, H.P., and Lovlie, R., p. 119–125, doi:10.1007/s003670050016. of a submarine slope, Tourelle Formation (Lower 1998, Glacigenic debris fl ows on the North Sea Trough Expedition 333 Scientists, 2011, NanTroSEIZE Stage 2: Ordovician), Quebec, Canada, in Doyle, L.J., and Mouth Fan during ice stream maxima: Marine Geol- Subduction inputs 2 and heat fl ow: Preliminary reports Pilkey, O.H., eds., Geology of continental slopes: Soci- ogy, v. 152, p. 217–246, doi:10.1016/S0025-3227 of the Integrated Ocean Drilling Program, v. 333, ety of Economic Paleontologists and Mineralogists (98)00072-3. doi:10.2204/iodp.pr.333.2011 Special Publication 27, p. 307–326. Kleverlaan, K., 1987, Gordo Megabed: A possible seis- Felix, M., and Peakall, J., 2006, Transformation of debris Hiscott, R.N., and Middleton, G.V., 1980, Fabric of coarse mite in a Tortonian submarine fan, Tabernas Basin, flows into turbidity currents: Mechanisms inferred deep-water sandstones, Tourelle Formation, Quebec, Province Almeria, southeast Spain: Sedimentary from laboratory experiments: Sedimentology, v. 53, Canada: Journal of Sedimentary Petrology, v. 50, Geology, v. 51, p. 165–180, doi:10.1016/0037-0738 p. 107–123, doi:10.1111/j.1365-3091.2005.00757.x. p. 703–722. (87)90047-9. Flemings, P.B., Behrmann, J.H., John, C.M., and Expedition Hodgson, D.M., 2009, Distribution and origin of hybrid beds Kneller, B.C., and Branney, M.J., 1995, Sustained high- 308 Scientists, 2006, Gulf of Mexico Hydrogeology: in sand-rich submarine fans of the Tanqua depocentre, density turbidity currents and the deposition of thick Proceedings of the Integrated Ocean Drilling Program, Karoo Basin, South Africa: Marine and Petroleum massive sands: Sedimentology, v. 42, p. 607–616, v. 308, doi:10.2204/iodp.proc.308.2006. Geology, v. 26, p. 1940–1957, doi:10.1016/j.marpetgeo doi:10.1111/j.1365-3091.1995.tb00395.x. Frenz, M., Wynn, R.B., Georgiopoulou, A., Bender, V.B., .2009.02.011. Kranenburg, C., and Winterwerp, J.C., 1997, Erosion of fl uid Hough, G., Masson, D.G., Talling, P.J., and Cronin, B.T., Hsu, S.K., Kuo, J., Lo, C.L., Tsai, C.H., Doo, W.B., Ku, mud layers. 1 Entrainment model: Journal of Hydraulic 2009, Provenance and pathways of late Quaternary turbi- C.Y., and Sibuet, J.C., 2008, Turbidity currents, sub- Engineering, v. 123, p. 504–511, doi:10.1061/ dites in the deep-water Agadir Basin, northwest African marine landslides and the 2006 Pingtung earthquake (ASCE)0733-9429(1997)123:6(504). margin: International Journal of Earth Sciences, v, 98, off SW Taiwan: Terrestrial Atmospheric and Oceanic Kuenen, P.H., 1951, Properties of turbidity currents of high no. 4, p. 721–733, doi.org/10.1007/s00531-008-0313-4. Science, v. 19, p. 767–772, doi:10.3319/TAO.2008 density, in Hough, J.L., ed., Turbidity currents and the Galy, V., France-Lanord, C., Beyssac, O., Faure, P., Kudrass, .19.6.767. transportation of coarse sediments to deep water: Society H., and Palhol, F., 2007, Effi cient organic carbon burial Huvenne, V.A.I., McPhail, S.D., Wynn, R.B., Furlong, of Economic Paleontologists and Mineralogists Special in the Bengal fan sustained by the Himalayan ero- M., and Stevenson, P., 2009, Mapping giant scours Publication 2, p. 14–33, doi:10.2110/pec.51.02.0014. sional system: Nature, v. 450, p. 407–410, doi:10.1038 in the deep ocean: Eos (Transactions, American Kuenen, P.H., 1966a, Experimental turbidite lamination in a /nature06273. Geophysical Union), v. 90, p. 274–275, doi:10.1029 circular fl ume: Journal of Geology, v. 74, p. 523–545, Gee, M.J.R., Masson, D.G., Watts, A.B., and Allen, P.A., /2009EO320002. doi:10.1086/627187. 1999, The Saharan debris fl ow: An insight into the Ilstad, T., Elverhøi, A., Issler, D., and Marr, J.G., 2004, Kuenen, P.H., 1966b, Matrix of turbidites: Experimen- mechanics of long run out submarine debris fl ow: Subaqueous debris fl ow behaviour and its dependence tal approach: Sedimentology, v. 7, p. 267–297, doi: Sedimentology, v. 46, p. 317–335, doi:10.1046/j.1365 on the sand/clay ratio: A laboratory study using par- 10.1111/j.1365-3091.1966.tb01295.x. -3091.1999.00215.x. ticle tracking: Marine Geology, v. 213, p. 415–438, Labaume, P., Mutti, E., and Seguret, M., 1987, Megaturbi- Gonzalez-Bonorino, G., and Middleton, G.V., 1976, A doi:10.1016/j.margeo.2004.10.017. dites—A depositional model from the Eocene of the Devonian submarine fan in western Argentina: Journal Ito, M., 2008, Downfan transformation from turbidity cur- SW-Pyrenean foreland basin, Spain: Geo-Marine Let- of Sedimentary Petrology, v. 46, p. 56–69. rents to debris fl ows at a channel-to-lobe transitional ters, v. 7, p. 91–101, doi:10.1007/BF02237988. Hafl idason, H., Lien, R., Sjerup, H.P., Forsberg, C.F., and zone: The lower Pleistocene Otadai Formation, Boso Laberg, J.S., and Vorren, T.O., 2000, Flow behaviour of Bryn, P., 2005, The dating and morphometry of the Peninsula, Japan: Journal of Sedimentary Research, the submarine glacigenic debris fl ows on the Bear Storegga Slide: Marine and Petroleum Geology, v. 22, v. 78, p. 668–682, doi:10.2110/jsr.2008.076. Island Trough Mouth Fan, western Barents Sea: Sedi- p. 123–136, doi:10.1016/j.marpetgeo.2004.10.008. Iverson, R.M., and Vallence, J.W., 2001, New views of granu- mentology, v. 47, p. 1105–1117, doi:10.1046/j.1365 Hampton, M.A., 1972, The role of subaqueous debris fl ow in lar mass fl ows: Geology, v. 29, p. 115–118, doi:10.1130 -3091.2000.00343.x. generating turbidity currents: Journal of Sedimentary /0091-7613(2001)029<0115:NVOGMF>2.0.CO;2. Lastras, G., DeBlasio, F.V., Canals, M., and Elverhøi, A., Petrology, v. 42, p. 775–793. Iverson, R.M., Logan, M., LaHusen, R.G., and Berti, M., 2005, Conceptual and numerical modelling of the Hampton, M.A., 1975, Competence of fi ne-grained debris 2010, The perfect debris fl ow? Aggregated results from BIG’95 debris fl ow, western Mediterranean Sea: Jour- flows: Journal of Sedimentary Petrology, v. 45, 28 large-scale experiments: Journal of Geophysical nal of Sedimentary Research, v. 75, p. 784–797, doi: p. 833–844. Research, v. 115, F03005, doi:10.1029/2009JF001514. 10.2110/jsr.2005.063.

486 Geosphere, June 2013

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/3/460/3343593/460.pdf by guest on 28 September 2021 Hybrid submarine sediment ﬂ ows

Locat, J., Lee, H.J., Nelson, C.H., Schwab, W.C., and Mohrig, D.G., and Marr, J.G., 2003, Constraining the effi - the margins of structurally confi ned submarine fans: Twichell, D.C., 1996, Analysis of the mobility of far ciency of turbidity current generation from submarine Marine and Petroleum Geology, v. 26, p. 1974–1996, reaching debris fl ows on the Mississippi Fan, Gulf of debris fl ows and slides using laboratory experiments: doi:10.1016/j.marpetgeo.2009.02.018. Mexico, in Senneset, K., ed., Proceedings of the Sev- Marine and Petroleum Geology, v. 20, p. 883–899, Revellino, P., Hungr, O., Guadagno, F.M., and Evans, S.G., enth International Symposium on Landslides, Trond- doi:10.1016/j.marpetgeo.2003.03.002. 2004, Velocity and runout simulation of destructive heim, Norway, Volume 1: Rotterdam, A.A. Balkema, Mohrig, D.G., Whipple, K., Ellis, C., and Parker, G., 1998, debris fl ows and debris avalanches in pyroclastic depos- p. 555–560. Hydroplaning of subaqueous debris fl ows: Geological its, Campania region, Italy: Engineering Geology, v. 45, Lowe, D.R., 1976, Subaqueous liquefi ed and fl uidized sedi- Society of America Bulletin, v. 110, p. 387–394, doi: p. 295–311, doi:10.1007/s00254-003-0885-z. ment fl ows and their deposits: Sedimentology, v. 23, 10.1130/0016-7606(1998)110<0387:HOSDF>2.3.CO;2. Ricci-Lucchi, F., and Valmori, E., 1980, Basin-wide turbi- p. 285–308, doi:10.1111/j.1365-3091.1976.tb00051.x. Mosher, D.C., and Piper, D.J.W., 2007, Analysis of multi- dites in a Miocene, over-supplied deep-sea plain: A Lowe, D.R., 1982, Sediment gravity fl ows; II, Deposi- beam seafl oor imagery of the Laurentian Fan and the geometrical analysis: Sedimentology, v. 27, p. 241– tional models with special reference to high density 1929 Grand Banks landslide area, in Lykousis, V., 270, doi:10.1111/j.1365-3091.1980.tb01177.x. turbidity currents: Journal of Sedimentary Petrology, et al., eds., Submarine mass movements and their con- Saller, A., Werner, K., Sugiaman, F., Cebastiant, A., May, v. 52, p. 279–298, doi:10.1306/212F7F31-2B24-11D7 sequences: 3rd International Symposium: Advances in R., Glenn, D., and Barker, C., 2008, Characteristics -8648000102C1865D. Natural and Technological Hazards Research 27: New of Pleistocene deep-water fan lobes and their applica- Lowe, D.R., and Guy, M., 2000, Slurry-fl ow deposits in the York, Springer, p. 77–88, doi:10.1007/978-1-4020 tion to an upper Miocene reservoir model, offshore Britannia Formation (Lower Cretaceous), North Sea: -6512-5_9. East Kalimantan Indonesia: American Association of A new perspective on the turbidity current and debris Mulder, T., and Alexander, A., 2001, The physical char- Petroleum Geologists Bulletin, v. 92, p. 919–949, doi: fl ow problem: Sedimentology, v. 47, p. 31–70, doi: acter of subaqueous sedimentary density fl ows and 10.1306/03310807110. 10.1046/j.1365-3091.2000.00276.x. their deposits: Sedimentology, v. 48, p. 269–299, doi: Schwab, W.C., Lee, H.J., Twichell, D.C., Locat, J., Nelson, Lowe, D.R., Guy, M., and Palfrey, A., 2003, Facies of 10.1046/j.1365-3091.2001.00360.x. C.H., McArthur, W.G., and Kenyon, N.H., 1996, Sedi- slurry-fl ow deposits, Britannia Formation (Lower Cre- Mulder, T., Savoye, B., and Syvitski, J.P.M., 1997, Numeri- ment mass-fl ow processes on a depositional lobe, outer taceous), North Sea: Implications for fl ow evolution cal modelling of a mid-sized gravity fl ow: The 1979 Mississippi Fan: Journal of Sedimentary Research, and deposit geometry: Sedimentology, v. 50, p. 45–80, Nice turbidity current (dynamics, processes, sediment v. 66, p. 916–927. doi:10.1046/j.1365-3091.2003.00507.x. budget and seafl oor impact): Sedimentology, v. 44, Shanmugam, G., and Moiola, R.J., 1995, Reinterpretation Macdonald, H.A., Wynn, R.B., Huvenne, V.A.I., Peakall, J., p. 305–326, doi:10.1111/j.1365-3091.1997.tb01526.x. of depositional processes in a classic fl ysch sequence Masson, D.G., Weaver, P.P.E., and McPhail, S.D., 2011, Mutti, E., 1992, Turbidite sandstones: AGIP, Istituto di Geo- (Pennsylvanian Jackfork Group), Ouachita Mountains, New insights into the morphology, fi ll, and remarkable logia, Università di Parma, 275 p. Arkansas and Oklahoma: American Association of longevity (>0.2 m.y.) of modern deep-water erosional Nelson, C.H., Twichell, D.C., Schwab, W.C., Lee, H.J., Petroleum Geologists Bulletin, v. 79, p. 672–695. scours along the northeast Atlantic margin: Geosphere, and Kenyon, N.H., 1992, Upper Pleistocene tur- Southard, J.B., 1991, Experimental determination of bed- v. 7, p. 845–867, doi:10.1130/GES00611.1. bidite sand beds and chaotic silt beds in the chan- form stability: Annual Review of Earth and Planetary Major, J.J., and Iverson, R.M., 1999, Debris-fl ow deposi- nelized, distal, outer-fan lobes of the Mississippi Sciences, v. 19, p. 423–455, doi:10.1146/annurev.ea tion: Effects of pore-fl uid pressure and friction concen- Fan: Geology, v. 20, p. 693–696, doi:10.1130/0091 .19.050191.002231. trated at fl ow margins: Geological Society of America -7613(1992)020<0693:UPTSBA>2.3.CO;2. Stow, D.A.V., and Bowen, A.J., 1980, A physical model Bulletin, v. 111, p. 1424–1434, doi:10.1130/0016 Normark, W.R., 1989, Observed parameters for turbid- for the transport and sorting of fi ne-grained sediment -7606(1999)111<1424:DFDEOP>2.3.CO;2. ity current fl ows in channels, Reservoir Fan, Lake by turbidity currents: Sedimentology, v. 27, p. 31–46, Marr, J.G., Harff, P.A., Shanmugam, G., and Parker, G., Superior: Journal of Sedimentary Petrology, v. 59, doi:10.1111/j.1365-3091.1980.tb01156.x. 2001, Experiments on subaqueous sandy gravity fl ows: p. 423–431. Sumner, E., Amy, L., and Talling, P.J., 2008, Deposit struc- The role of clay and water content in fl ow dynamics and Normark, W.R., Posamentier, H., and Mutti, E., 1993, ture and processes of sand deposition from a decel- depositional structures: Geological Society of America Turbidite systems: State of the art and future direc- erating sediment suspension: Journal of Sedimentary Bulletin, v. 113, p. 1377–1386, doi:10.1130/0016 tions: Reviews of Geophysics, v. 31, p. 91–116, doi: Research, v. 78, p. 529–547, doi:10.2110/jsr.2008.062. -7606(2001)113<1377:EOSSGF>2.0.CO;2. 10.1029/93RG0283. Sumner, E.J., Talling, P.J., and Amy, L.A., 2009, The depos- Masson, D.G., Huggett, Q.J., and Brunsden, D., 1993, The Normark, W.R., Piper, D.J.W., Posamentier, H., Pirmez, its of fl ows transitional between turbidity currents and surface texture of the Saharan debris fl ow and some C., and Migeon, S., 2002, Variability in form and debris fl ow: Geology, v. 37, p. 991–994, doi:10.1130 speculations on submarine debris flow processes: growth of sediment waves on turbidite channel levees: /G30059A.1. Sedimentology, v. 40, p. 583–598, doi:10.1111/j.1365 Marine Geology, v. 192, p. 23–58, doi:10.1016/S0025 Sumner, E.J., Talling, P.J., Amy, L.A., Wynn, R.B., Steven- -3091.1993.tb01351.x. -3227(02)00548-0. son, C., and Frenz, M., 2012, Facies architecture of indi- Masson, D.G., Van Niel, B., and Weaver, P.P.E., 1997, Flow Normark, W.R., Piper, D.J.W., and Sliter, R., 2006, Sea-level vidual basin-plain turbidites: Comparison with existing processes and sediment deformation in the Canary and tectonic control of middle to late Pleistocene tur- models and implications for fl ow processes: Sedimen- debris fl ow on the NW African continental rise: Sedi- bidite systems in Santa Monica Basin, offshore Cali- tology, v. 59, p. 1850–1857, doi:10.1111/j.1365-3091 mentary Geology, v. 110, p. 163–179, doi:10.1016 fornia: Sedimentology, v. 53, p. 867–897, doi:10.1111 .2012.01329.x. /S0037-0738(96)00089-9. /j.1365-3091.2006.00797.x. Sylvester, Z., and Lowe, D.R., 2004, Textural trends in turbi- Masson, D.G., Wynn, R.B., and Talling, P.J., 2010, Large Parchure, T.M., and Mehta, A.J., 1985, Erosion of soft cohe- dites and slurry beds from the Oligocene fl ysch of the landslides on passive continental margins: Processes, sive sediment deposits: Journal of Hydraulic Engineer- Carpathians, Romania: Sedimentology, v. 51, p. 945– hypotheses and outstanding questions, in Mosher, ing, v. 111, p. 1308–1326, doi:10.1061/(ASCE)0733 972, doi:10.1111/j.1365-3091.2004.00653.x. D.C., et al., eds., Submarine mass movements and their -9429(1985)111:10(1308). Talling, P.J., Peakall, J., Sparks, R.S.J., Ó Cofaigh, C.S., consequences: 4th International Symposium: Advances Payros, A., Pujalte, V., and Orue-Etxebarria, X., 1999, Dowdeswell, J.A., Felix, M., Wynn, R.B., Baas, J.H., in Natural and Technological Hazards Research 28: The South Pyrenean Eocene carbonate megabreccias Hogg, A.J., Masson, D.G., Taylor, J., and Weaver, Dordrecht, Springer, p. 153–166, doi:10.1007/978-90 revisited: New interpretation based on evidence from P.P.E., 2002, Experimental constraints on shear mix- -481-3071-9. the Pamplona Basin: Sedimentary Geology, v. 125, ing rates and processes: Implications for the dilution McAnally, W.H., Friedrichs, C., Hamilton, D., Hayter, E., p. 165–194, doi:10.1016/S0037-0738(99)00004-4. of submarine debris fl ows, in Dowdeswell, J.A., and Shrestha, P., Rodriguez, H., Sheremet, A., and Teeter , Pierson, T.C., Janda, R.J., Thouret, J.C., and Borrera, C.A., O’Cofaigh, C.S., eds., Glacier-infl uenced sedimentation A., 2007, Management of fl uid mud in estuaries, 1990, Perturbation and melting of snow and ice by the on high-latitude continental margins: Geological Soci- bays, and lakes. I: Present state of understanding on 13 November 1985 eruption of Nevado del Ruiz, Colom- ety of London Special Publication 203, p. 89–103, doi: character and behaviour: Journal of Hydraulic Engi- bia, and consequent mobilization, fl ow and deposition of 10.1144/GSL.SP.2002.203.01.06. neering, v. 133, p. 9–22, doi:10.1061/(ASCE)0733 lahars: Journal of Volcanology and Geothermal Research, Talling, P. J., Amy, L. A., Wynn, R. B., Peakall, J., and Rob- -9429(2007)133:1(9). v. 41, p. 17–66, doi:10.1016/0377-0273(90)90082-Q. inson, M., 2004, Beds comprising debrite sandwiched McCaffrey, W., and Kneller, B., 2001, Process controls on Piper, D.J.W., and Savoye, B., 1993, Processes of Late within co-genetic turbidite: Origin and widespread the development of stratigraphic trap potential on the Quaternary turbidity-current flow and deposition occurrence in distal depositional environments: Sedi- margins of confi ned turbidite systems and aids to res- on the Var deep-sea fan, north-west Mediterranean mentology, v. 51, p. 163–194. ervoir evaluation: American Association of Petroleum Sea: Sedimentology, v. 40, p. 557–582, doi:10.1111 Talling, P.J., Wynn, R.B., Masson, D.G., Frenz, M., Cronin, Geologists Bulletin, v. 85, p. 971–988. /j.1365-3091.1993.tb01350.x. B.T., Schiebel, R., Akhmetzhanov, A.M., Dallmeier- McCave, I.N., and Jones, K.P.N., 1988, Deposition of Piper, D.J.W., Cochonat, P., and Morrison, M.L., 1999, The Tiessen, S., Benetti, S., Weaver, P.P.E., Georgiopou- ungraded muds from high-density non-turbulent tur- sequence of events around the epicentre of the 1929 lou, A., Zühlsdorff, C., and Amy, L.A., 2007a, Onset bidity currents: Nature, v. 333, p. 250–252, doi:10.1038 Grand Banks earthquake: Initiation of debris fl ows of submarine debris fl ow deposition far from origi- /333250a0. and turbidity currents inferred from sidescan sonar: nal giant landslide: Nature, v. 450, p. 541–544, doi: Migeon, S., Ducassou, E., Le Gonidec, Y., Rouillard, P., Sedimentology, v. 46, p. 79–97, doi:10.1046/j.1365 10.1038/nature06313. Mascle, J., and Revel-Rolland, M., 2010, Lobe con- -3091.1999.00204.x. Talling, P.J., Amy, L.A., and Wynn, R.B., 2007b, New struction and sand/mud segregation by turbidity cur- Pyles, D.R., and Jennette, D.C., 2009, Geometry and archi- insights into the evolution of large volume turbidity rents and debris fl ows on the western Nile deep-sea tectural associations of co-genetic debrite-turbidite currents; comparison of turbidite shape and previous fan (Eastern Mediterranean): Sedimentology, v. 229, beds in basin-margin strata, Carboniferous Ross Sand- modelling results: Sedimentology, v. 54, p. 737–769, p. 124–143, doi:10.1016/j.sedgeo.2010.02.011. stone (Ireland): Applications to reservoirs located on doi:10.1111/j.1365-3091.2007.00858.x.

Geosphere, June 2013 487

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/3/460/3343593/460.pdf by guest on 28 September 2021 Talling

Talling, P.J., Amy, L.A., Wynn, R.B., Blackbourn, G., and ern North American margin: Sedimentology, v. 55, ling, G., eds., Sedimentation in submarine canyons, Gibson, O., 2007c, Turbidity current evolution deduced p. 97–136, doi:10.1111/j.1365-3091.2007.00894.x. fans and trenches: London, Dowden, Hutchinson and from extensive thin turbidites: Marnoso-arenacea Twichell, D.C., Schwab, W.C., Nelson, C.H., Kenyon, N.H., Ross, Inc., p. 190–109. Formation (Miocene), Italian Apennines: Journal of and Lee, H.J., 1992, Characteristics of a sandy deposi- Weaver, P.P.E., Masson, D.G., Gunn, D.E., Kidd, R.B., Sedimentary Research, v. 77, p. 172–196, doi:10.2110 tional lobe on the outer Mississippi fan from SeaMARC and Maddison, D.A., 1994, Sediment mass wasting /jsr.2007.018. 1A sidescan sonar images: Geology, v. 20, p. 689–692, in the Canary Basin, in Pickering, K.T., et al., eds., Talling, P.J., Wynn, R.B., Rixon, R., Schmidt, D., Sumner, doi:10.1130/0091-7613(1992)020<0689:COASDL>2.3 Atlas of deep-water environments: Architectural style E., and Amy, L.A., 2010, How did submarine fl ows .CO;2. in turbidite systems: London, Chapman and Hall, transport boulder sized mud clasts to the fringes of the Twichell, D.C., Schwab, W.C., and Kenyon, N.H., 1995, p. 287–296. Mississippi Fan?: Journal of Sedimentary Research, Geometry of sandy deposits at the distal edge of the Winterwerp, J.C., and Kranenburg, C., 1997, Erosion of v. 80, p. 829–851, doi:10.2110/jsr.2010.076. Mississippi Fan, Gulf of Mexico, in Pickering, K.T., fluid mud layers. 2 Experiments and model vali- Talling, P.J., Sumner, E.J., Masson, D.G., and Malgesini, et al., eds., Atlas of deep-water environments: Archi- dation: Journal of Hydraulic Engineering, v. 123, G., 2012a, Subaqueous sediment density fl ows: Depo- tectural style in turbidite systems: London, Chapman p. 512–519, doi:10.1061/(ASCE)0733-9429(1997)123:6 sitional processes and deposit types: Sedimentology, & Hall, p. 282–286. (512). v. 59, p. 1937–2003, doi:10.1111/j.1365-3091.2012 Twichell, D.C., Schwab, W.C., Kenyon, N.H., and Lee, Wood, A., and Smith, A.J., 1958, The sedimentation and .01353.x. H.J., 1996, Breaching the levee of a channel on the sedimentary history of the Aberystwyth Grits (upper Talling, P.J., Malgesini, G., Sumner, E.J., Amy, L.A., Felletti, Mississippi Fan, in Gardner, J.V., et al., eds., Geol- Llandoverian): Geological Society of London Quar- F., Blackbourn, G., Nutt, C., Wilcox, C., Harding, I.C., ogy of the United States’ seafloor: The view from terly Journal, v. 114, p. 163–195, doi:10.1144/gsjgs and Akbari, S., 2012b, Planform geometry, stacking pat- GLORIA : Cambridge, U.K., Cambridge University .114.1.0163. tern, and extrabasinal origin of low strength and inter- Press, p. 85–96. Wright, S., and Parker, G., 2004, Density stratifi cation effects mediate strength cohesive debris fl ow deposits in the Twichell, D., Nelson, C.H., Kenyon, N., and Schwab, W., in sand-bed rivers: Journal of Hydraulic Engineering, Marnoso-arenacea Formation, Italy: Geosphere, v. 8, 2009, The infl uence of external processes on the Holo- v. 130, p. 783–795, doi:10.1061/(ASCE)0733-9429 p. 1207–1230, doi:10.1130/GES00734.1. cene evolution of the Mississippi Fan, in Kneller, B., (2004)130:8(783). Talling, P.J., Malgesini, G., and Felletti F., 2012c, Can lique- et al., eds., External controls on deepwater depositional Wynn, R.B., Weaver, P.P.E., Masson, D.G., and Stow, D.A.V., fi ed submarine debris fl ows deposit clean sandstone systems: SEPM (Society for Sedimentary Geology) 2002, Turbidite depositional architecture across three over large areas? Field evidence from the Marnoso- Special Publication 92, p. 145–159. inter-connected deep-water basins on the Northwest Arenacea Formation, Italian Apennines: Sedimentol- Vangriesheim, A., Khripounoff, A., and Crassous, P., 2009, African Margin: Sedimentology, v. 49, p. 669–695, ogy, doi:10.1111/j.1365–3091.2012.01358.x. Turbidity events observed in situ along the Congo sub- doi:10.1046/j.1365-3091.2002.00471.x. Tripsanas, E.K., and Piper, D.J.W., 2008, Glaciogenic marine channel. Deep-Sea Research II, v. 56, p. 2208– Wynn, R.B., Talling, P.J., Masson, D.G., Stevenson, C.J., debris-fl ow deposits of Orphan Basin, offshore eastern 2222, doi:10.1016/j.dsr2.2009.04.004. Cronin, B.T., and Le Bas, T.P., 2010, Investigating the Canada: Sedimentological and rheological proper- Van Maren, D.S., Winterwerp, J.C., Wang, Z.Y., and Pu, timing, processes and deposits of one of the world’s ties, origin, and relationship to meltwater discharge: Q., 2009, Suspended sediment dynamics and morpho- largest submarine gravity fl ows: The ‘Bed 5 event’ Journal of Sedimentary Research, v. 78, p. 724–744, dynamics in the Yellow River, China: Sedimentology, off northwest Africa in Mosher, D.C., et al., eds., doi:10.2110/jsr.2008.082. v. 56, p. 785–806, doi:10.1111/j.1365-3091.2008 Submarine mass movements and their consequences: Tripsanas, E.K., Piper, D.J.W., Jenner, K.A., and Bryant, .00997.x. 4th International Symposium: Advances in Natural W.R., 2008, Submarine mass-transport facies: New Van Vliet, A., 1978, Early Tertiary deepwater fans of and Technological Hazards Research 28: Dordrecht, perspectives on fl ow processes from cores on the east- Guipuzcoa, northern Spain, in Stanley, D.J., and Kel- Springer, p. 463–474.

488 Geosphere, June 2013

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/9/3/460/3343593/460.pdf by guest on 28 September 2021