Sedimentary Geology 179

www.elsevier.com/locate/sedgeo

Sedimentary Geology

Editorial

Sediment gravity flows: Recent advances in process and field

analysis—introduction

A popular axiom reads: bWe know more about the

surface of the moon than we do about the ocean floorQ.Although the legitimacy of this saying is debatable, as

it compares apples and oranges (Are the bdeadQ moon

and the hydrodynamically and biologically active

ocean floor fair opponents?), there is certainly truth

in the connotation that the deep ocean is one of the

last frontiers of human exploration on Earth. In the

last three decades significant progress has been made

in our understanding of sedimentary processes and

environments in the deep sea. Progress has accelerated

in recent years, driven by social and economic

motives in conjunction with continuing scientific

curiosity. Deep-marine sedimentary rocks store the

world’s largest economic reserves of oil and gas, and

near-surface gas hydrates are a potential future source

of energy. The risk of marine geohazards, made all too

clear by the devastating tsunami in the Indian Ocean

(December 26th 2004), is another incentive for

conducting oceanic research. Typically catastrophic

sediment gravity flows, such as turbidity currents,

debris flows, slides and slumps, are a high risk factor

for man-made engineering works in deep water. These

include communication cables and drilling rigs whose

number, on the continental slope, has multiplied with

the shift of hydrocarbon exploration targets to deeper

water. The deep sea is also one of the most important

storage reservoirs of (palaeo)climate signals, e.g.

through stable isotopes and microfossils. Last but

not least, the desire to understand the fundamental

dynamics of the dispersal of sediment from the

continent via the shelf onto the continental slope

0037-0738/$ - see front matter D 2005 Elsevier B.V. All rights reserved.

doi:10.1016/j.sedgeo.2005.05.003

and into the abyss has grasped academia ever since

the pioneering work of Philip Kuenen, Gerry Mid-

dleton, Emiliano Mutti, Franco Ricci Lucchi, Roger

Walker and others.

Despite the progress made in recent years, our

knowledge of deep-marine sedimentary environments

is still sketchy compared to other environments, and it

is largely based on conceptual models for which

validation is pending. However, new exciting techni-

ques that hold promise for filling these knowledge

gaps have been developed in the various strands of

deep-marine sedimentary geology. Sea-going research

has the benefit of exploration tools with ever

increasing resolution (e.g. giant coring, multibeam

sonar and 3D seismic systems), thus starting to bridge

the gap to scales of observation typical of outcrops of

deep-marine sedimentary rocks. In turn, outcrop

studies have won ground by focussing on well-

exposed seismic-scale sedimentary sequences, such

as in the Karoo Basin (South Africa), East Greenland

and Pakistan. Digital elevation mapping is a promis-

ing new technique in outcrop work, particularly when

considered as an add-on to more classic field work. In

laboratory-scale research of sediment gravity flows

and their deposits, it is now possible to quantify

physical parameters within sediment gravity flows of

high density, using, for example, Ultrasonic Doppler

Velocity methods to measure flow velocity and Ultra

High Concentration Meters to measure concentration

of suspended particles. Before, such data were

obtainable only at the free boundaries of such flows

and in flows of very low density. Moreover, labo-

179 (2005) 1–3

pc

STAMPA KAI TELEIA

Editorial2

ratory experiments have potential to act as intermedi-

ate between advanced numerical models of deep-

marine sediment transport and real processes in

nature, because, at the moment, validation of numer-

ical models is difficult due to the lack of suitable

natural datasets.

This special issue is a collection of papers

presented at the annual general meeting of the British

Sedimentologists Research Group (BSRG) in Decem-

ber 2003 at the School of Earth and Environment,

University of Leeds, United Kingdom. The papers

investigate sediment gravity flows and deposits, but

from different perspectives and with different meth-

ods. As such, they provide insides into recent

advances in deep-marine sedimentology, following

some of the trends mentioned above. The special issue

starts with three papers that approach the flow

dynamics of sediment gravity flows from an exper-

imental perspective, thus studying scale models of

these flows in laboratory flumes. Amy, Peakall et al.

studied the temporal and spatial evolution of density

currents with imposed bipartite vertical stratification

of flow density and viscosity. It was found that the

lower flow layer outruns or lags behind the upper flow

layer, depending on the density and viscosity contrast

between these layers. This may have important

implications for the depositional signature of stratified

sediment gravity flows in nature, as shown by field

examples. Choux et al. analyse the spatio-temporal

evolution of downstream velocity, turbulence inten-

sity, median grain size and suspended sediment

concentration of two laboratory-scale turbidity cur-

rents of different initial density. Links between

internal fluid and sediment dynamics are proposed,

and dimensionless variables are developed with the

aim to predict the dynamic behaviour of particulate

gravity flows across a wider range of concentrations.

The flows investigated by Choux et al. were low

density, non-cohesive and depositional. This is in

contrast to the experimental particle-laden gravity

currents studied by Felix et al., which were low- and

high-density, cohesive and non-cohesive, and non-

depositional. One of their conclusions is that vertical

flow regions of high turbulence intensity in low-

density, non-cohesive turbidity currents are coupled

through turbulent mixing, while similar regions in

high-density flows are decoupled due to turbulence

suppression. In contrast, however, Choux et al. found

that coherent flow structures generated at the lower

and upper boundaries of their turbidity currents were

discontinuous across the interjacent level of maximum

velocity. Further research is required to explain these

apparently conflicting observations. Detailed analysis

of differences in methodology and flow type (e.g.

depositional versus non-depositional flow) may well

be a proper basis for such work. None of the three

experimental papers provide information on deposits

formed by the laboratory flows. Yet, Amy, Peakall et

al. and Felix et al. included a discussion on the

implications for the geometry and internal organisa-

tion of natural deposits. At present these implications

are rather speculative. Future work should therefore

concentrate on the properties of laboratory deposits as

much-needed intermediate step between scale models

of flow dynamics and prototypes of natural deposi-

tional products.

The two papers that follow the experimental papers

present marine geological data of recent sediment

gravity flow deposits using state-of-the-art geophys-

ical instrumentation. Cronin et al. investigated the

morphology and fill history of Donegal Bay sub-

marine canyon in the Rockall Trough (Irish Sea) using

side scan sonar of different frequencies, 3.5 kHz echo

sounding, multibeam seismic profiling and gravity

coring. Changes in the fill history of the canyon are

closely related to sea level rise since the last

glaciation, with active transport of coarse-grained

sediment during low sea level and infrequent transport

of finer-grained sediment during high sea level. In a

well-illustrated paper, Cunningham et al. evaluate

along- and down-slope sedimentary processes along

the Celtic Margin outer shelf and upper slope of the

Bay of Biscay between Goban Spur and Brenot Spur,

using multi-beam bathymetry and backscatter, 3.5

kHz echo sounding, side-scan sonar and seabed

samples. They reconstruct sediment transport path-

ways from the spatial orientation of giant sandwaves,

and demonstrate that faulting has played an important

role in canyon development and in the initiation of

slope failure events as turbidity currents.

The final three papers of this special issue take the

classic, well-tested approach of outcrop sedimentol-

ogy to reconstruct flow dynamics and sediment

dispersal patterns from deep-marine sedimentary

facies. Edwards et al. focussed their attention on the

Turonian Mancos Shale in the Western Interior

Editorial 3

foreland basin of Utah, USA, where channellised

turbidity current deposits are interpreted as the

product of hyperpycnal flows. Sequence-stratigraphic

concepts are used to explain the occurrence of the

hyperpycnal flows. Wynn et al. introduce a new

technique, called digiscoping, to obtain sedimento-

logical data from inaccessible outcrops. Digiscoping

combines the high resolution of modern digital

cameras with the high magnification of field tele-

scopes to resolve features on the scale of centimetres

from distances of several hundred metres, as exem-

plified by a study of the Fontanelice Channels in the

Marnoso Arenacea Formation, Italian Apennines. By

using extensive correlations of deep-marine deposits

in the same formation, Amy, Talling et al. were able to

distinguish between thick-bedded sandstones of tur-

biditic and debris-flow origin. Principal recognition

criteria are based on kilometre-scale changes in the

geometry of individual beds and submetre-scale

structural and textural properties. Amy, Talling et

al.’s work nicely illustrates the above-mentioned trend

towards studies of large-scale, well-exposed deep-

marine sedimentary successions, yet the authors also

fully realise the value of smaller-scale observations

for process-based interpretations. This approach may

well contribute to bridging the gap between typical

field-scale and seismic-scale observations.

Acknowledgments

This special issue would not have been possible

without the help of more than 20 reviewers. On behalf

of the authors, I thank all of them for their time and

effort devoted to improving the quality of the papers.

The editorial support of Maarten Felix is greatly

appreciated. Tirza Van Daalen and Tonny Smit (both

Elsevier) provided welcome logistical advice.

Jaco H. Baas

Earth Sciences, School of Earth and Environment,

University of Leeds, Leeds LS2 9JT, United Kingdom

E-mail address: [email protected].

Tel.: +44 113 3436624; fax: +44 113 3435259.


Sedimentary Geology

Density- and viscosity-stratified gravity currents: Insight from

laboratory experiments and implications for submarine

flow deposits

L.A. Amy a,b,*, J. Peakall b, P.J. Talling a

aCentre for Environmental and Geophysical flows, Department of Earth Sciences, University of Bristol, Bristol, BS8 1RJ, UKbEarth and Biosphere Institute School of Earth and Environment, University of Leeds, Leeds, LS2 9JT, UK

Received 22 April 2004; accepted 6 April 2005

Abstract

Vertical stratification of particle concentration is a common if not ubiquitous feature of submarine particulate gravity flows.

To investigate the control of stratification on current behaviour, analogue stratified flows were studied using laboratory

experiments. Stratified density currents were generated by releasing two-layer glycerol solutions into a tank of water. Flows

were sustained for periods of tens of seconds and their velocity and concentration measured. In a set of experiments the strength

of the initial density and viscosity stratification was increased by progressively varying the lower-layer concentration, CL. Two

types of current were observed indicating two regimes of behaviour. Currents with a faster-moving high-concentration basal

region that outran the upper layer were produced if CLb75%. Above this critical value of CL, currents were formed with a

relatively slow, high-concentration base that lagged behind the flow front. The observed transition in behaviour is interpreted to

indicate a change from inertia- to viscosity-dominated flow with increasing concentration. The reduction in lower-layer velocity

at high concentrations is explained by enhanced drag at low Reynolds numbers. Results show that vertical stratification

produces longitudinal stratification in the currents. Furthermore, different vertical and temporal velocity and concentration

profiles characterise the observed flow types. Implications for the deposit character of particle-laden currents are discussed and

illustrated using examples from ancient turbidite systems.

D 2005 Elsevier B.V. All rights reserved.

Keywords: Sediment gravity flows; Turbidity currents; Subaqueous debris flows; Flow stratification; Analogue experiments


doi:10.1016/j.sedgeo.2005.04.009

* Corresponding author. Centre for Environmental and Geophy-

sical flows, Department of Earth Sciences, University of Bristol,

Bristol, BS8 1RJ, UK. Tel.: +44 117 954 5235; fax: +44 117 925

3385.

E-mail address: [email protected] (L.A. Amy).

1. Introduction

Sediment-laden density flows with a wide range of

sediment concentrations occur in subaqueous environ-

ments (Mulder and Alexander, 2001). These density

flows include turbidity currents and debris flows.

179 (2005) 5–29

L.A. Amy et al. / Sedimentary Geology 179 (2005) 5–296

Collectively they dominate sediment flux from shal-

low to deep marine environments in many locations

(Kneller and Buckee, 2000), and form some of the

most voluminous sediment accumulations on Earth

(Bouma et al., 1985). Individual events may transport

tens, or even hundreds, of cubic kilometres of sedi-

ment (Piper et al., 1999; Wynn et al., 2002). Ancient

deposits of sedimentary density flows form many of

the world’s largest petroleum reservoirs (Weimer and

Link, 1991), whilst modern flow events are a signif-

icant hazard to seafloor structures (Barley, 1999).

Almost all particulate gravity flows contain verti-

cal gradients in suspended-sediment concentration

where particle concentration decreases upwards

away from the bed (Fig. 1). Particle stratification is

a property of laboratory currents (e.g., Middleton,

1966; Postma et al., 1988), numerical simulations

(e.g., Stacey and Bowen, 1988; Felix, 2002), and

is recorded in the few natural flows that have been

instrumented (e.g., Normark, 1989; Chikita, 1990).

This characteristic develops within flows carrying a

range of grain sizes since it takes more energy to

suspend relatively dense and large particles at a point

above the bed. Stratification may become more pro-

nounced due to variable rates of particle settling,

BED

Mixing at upperflow boundary

Particle settling

Particleentrainment

Differentialsupport ofgrains withvarying mass

BED

Fig. 1. Schematic diagram showing particle stratification that occurs

within sediment gravity flows and its principal causes.

mixing with ambient fluid and entrainment of sub-

strate material (Fisher, 1995; Peakall et al., 2000;

Gladstone et al., 2004). Since both the density and

viscosity of particle–fluid mixtures are governed by

particle concentration, flows are also vertically strati-

fied in terms of these properties.

There is strong evidence for stratification in par-

ticulate density currents, however, few studies have

investigated in a systematic manner how density and

viscosity stratification influence flow behaviour.

Gladstone et al. (2004) investigated the behaviour of

two-layer, lock-exchange, stratified density flows

using laboratory experiments. They demonstrated

that layer density and volume have a marked effect

on the current’s evolution and the resulting flow

structure. Their results are applicable to inertial,

surge-type density-stratified currents. In this study

experiments were run to investigate currents that

were viscosity-stratified and density-stratified with

relatively long durations. These experimental currents

should be more representative of natural particle-laden

currents with relatively large volumes and high parti-

cle concentration (Peakall et al., 2001). A series of

experiments is presented in which the initial density

and viscosity stratification of solute-driven currents

was systematically varied. In these experiments the

velocity and concentration structure of flows were

recorded using instrumentation. The interaction be-

tween layers was also analysed from recorded video

footage. These experiments allow the role of density

and viscosity stratification on the behaviour of the flow

to be assessed. Implications for sediment deposition

from particle-laden currents and resulting deposit char-

acter are discussed.

2. Flow stratification

This paper presents new data on the concentration

distributions of density currents. Existing data on the

stratification of sediment gravity flows has been

reviewed by Peakall et al. (2000) and Kneller and

Buckee (2000). Laboratory studies have shown that

different types of flow stratification can occur

depending on flow concentration. Relatively low

concentration (b10% by volume) and fully turbulent,

depositional currents display broadly continuous pro-

files with concentration decreasing gradually up-

L.A. Amy et al. / Sedimentary Geology 179 (2005) 5–29 7

wards (Fig. 2A–C). A gradual concentration profile

occurs under both subcritical and supercritical flow

conditions. Grain-size classes are distributed differ-

ently throughout the flow depth (Garcıa, 1994). Finer

particles tend to be more evenly distributed through-

out the flow depth compared to coarser grains. Nu-

merical models (Stacey and Bowen, 1988; Felix,

2002) and measurements of natural turbidity currents

(Normark, 1989; Chikita, 1990) also suggest contin-

A.

D.

C.

E. F.

Coarse

Fine

Hei

ght a

bove

bed

Concentrationor velocity

VelocityConcentration

B.

Fig. 2. Measured concentration and velocity profiles of laboratory

sediment gravity flows. The vertical dimension is normalised with

respect to the height of the velocity maximum and the horizontal

scale is normalised using the velocity and concentration maxima.

(A) Continuous concentration profile; strongly depositional subcrit-

ical turbidity current (Garcıa, 1994). The distributions of fine (5 Am)

and coarse (32 Am) grain size fractions are also shown. (B) Nearly

continuous concentration profile (slight inflexion above the velocity

maximum); weakly depositional subcritical turbidity current on a

low-angle slope (Altinakar et al., 1996). (C) Nearly continuous

concentration profile; low-concentration (1055 kg m�3) fluid mud-

flow (van Kessel and Kranenburg, 1996). (D) Two-layer model with

a stepped concentration profile above the velocity maximum, based

on visual observations of strongly depositional lock release turbidity

currents (Middleton, 1966, 1993). (E) Stepped concentration and

velocity profile; high-concentration (1200 kg m�3) fluid mudflow

(van Kessel and Kranenburg, 1996). (F) Multi-stepped concentra-

tion profile inferred from video and modified velocity profile mea-

sured using trajectories of moving particles; high-concentration

turbidity current with starting concentration of 35–40% volume

fraction (Postma et al., 1988).

uous profiles for relatively low-concentration cur-

rents that are weakly depositional.

Few measurements of the concentration profiles

of high-concentration, particulate laboratory currents

exist (N20% by volume). Based on visual observa-

tions of strongly depositional currents, Middleton

(1966, 1993) proposed a two-layer model. This

model suggests a stepped profile with a high-con-

centration lower layer overlain by a more dilute and

relatively turbulent upper layer, also observed in

other experiments (e.g., Hampton, 1972; Mohrig et

al., 1998; Hallworth and Huppert, 1988; Marr et al.,

2001). A study by van Kessel and Kranenburg

(1996) measured the concentration profiles of fluid

mudflows and demonstrated a stepped profile for

those with relatively high concentrations. More

importantly, in a set of experiments, they were able

to document a change from a gradual (Fig. 2C) to a

stepped profile (Fig. 2E) with increasing mud con-

centration and were able to show that this occurred

in conjunction with a transition from turbulent to

laminar flow. In another experimental study on a

high-concentration flow, using cohesionless sedi-

ment, a laminar-moving high-concentration basal

layer was observed to form (Postma et al., 1988).

This took the form of a wedge propagating behind

the head of the current and below an overriding

turbulent, but strongly stratified, upper region. The

inferred velocity and concentration profile in these

currents is slightly different to those of high-concen-

tration fluid mudflows (Fig. 2F). The velocity profile

has an doverhanging noseT with a discrete reduction

in values below the velocity maximum, whilst the

concentration profile has two steps one below and

one above the velocity maximum. However, these

measurements were estimated somewhat crudely

compared to the other reported studies, because

flow velocity was derived from the motion of parti-

cles adjacent to the tank wall captured by film, and

thus do not exclude wall affects, whilst concentration

was estimated visually.

3. Experimental method

The experiments were run in a glass-walled, grav-

ity-current tank located in the School of Earth and

Environment, University of Leeds. The tank was 6 m


in length, 0.5 m in width and was filled with water

to a depth of ~1.5 m (Fig. 3). Stratified flows were

generated using two aqueous solutions with different

concentrations of glycerol. Each solution was first

mixed in an external reservoir tank and circulated

from this tank into a header box and back again via

an overflow system. Solutions were dyed different

colours and seeded with small amounts (b1% by

volume) of millimetre-sized neutrally buoyant parti-

cles to aid flow visualisation. A small amount of

silica flour (grain size b50 Am diameter) was also

added to help flow velocity measurement (see

below). Experiments were started by opening a

valve on the header boxes and allowing the solutions

to gravity drain into the main tank. The discharge

rate from each header box was kept constant by

maintaining a constant fluid head in the header

box. It was not possible, however, to keep the dis-

charge rate the same for solutions of different gly-

cerol concentrations. The discharge rate varied for

fluids of different glycerol concentrations (i.e., be-

tween different layers and experiments) from 2–4 l/s

owing to their different fluid densities and viscosi-

ties. The solutions passed through an inlet box which

was laid flat on the tank floor. The inlet box parti-

tioned the two glycerol solutions into a vertically

stratified release, with a lower relatively dense

layer and an upper less-dense layer. It also dampened

the initial turbulence by passing the fluid through a

Fig. 3. The experimental set-up used. The apparatus consists of a gravity cu

shown for clarity and is not drawn to scale.

section filled with polystyrene chips. Currents flowed

down a 3.5-m-long, smooth, floor inclined at 38.Fluid at the end of the tank was collected in a

sump and pumped out of the tank to minimise the

effects of flow reflection and changes in the ambient

fluid depth.

3.1. Solute properties

In these experiments solutions were used instead of

particle–water slurries. This approach was chosen

since it was technically difficult to generate high-

concentration currents with long durations and with

controlled initial stratifications using sediment. Solu-

tions of aqueous glycerol were chosen since they have

similarities in their density and viscosity with sedi-

ment–water mixtures. Glycerol (C3H8O3) has a den-

sity of 1260 kg m�3 and viscosity of ~1.5 kg m�1 s�1

at 20 8C (CRC Handbook of Chemistry and Physics).

For aqueous glycerol solutions viscosity increases by

several orders of magnitude with increasing glycerol

concentrations from ~10�3 to 1.5 kg m�1 s�1 (Fig.

4). The viscosity of sediment–water mixtures also

increases strongly by several orders of magnitude at

particle concentrations greater than 40–50% for mix-

tures composed of cohesionless particles (Richardson

and Zaki, 1954; Kreiger and Dougherty, 1959), and at

relatively low particle concentrations for those con-

taining cohesive particles (Major and Pierson, 1992;

rrent tank and two external reservoir tanks. Only one reservoir tank is

0 20 40 60 80 100

Glycerol concentration, % by weight

0.001

0.01

0.1

1

10

Vis

cosi

ty,k

gm

-1s-

1

0 10 20 30 40 50 60

Particle concentration, % by volume

Fluid mixtureAqueous glycerol solution (CRC Handbook)Kaolinite / china clay (Dewit, 1992)Cohesionless particles (Krieger and Dougherty, 1959)Silicon carbide (Ferreira and Diz, 1999)

Fig. 4. The relationship between concentration and viscosity for

glycerol solutions and several types of particle–water mixtures. The

data shown for aqueous glycerol mixtures and china clay bearing

mixtures are based on empirical data from the CRC Handbook of

Chemistry and Physics and from De Wit (1992), respectively. The

relationship for cohesionless mixtures is taken from the theoretical

model proposed by Kreiger and Dougherty (1959) for hard spheres

suspended in water. The relationship for silicon carbide mixtures is

a modified Kreiger and Dougherty model fitted to experimental data

of slurries containing particles with a mean particle size of 13 Amand measured at shear rates of ~100 s�1 (Ferreira and Diz, 1999).


Coussot, 1997). Density increases linearly with con-

centration for both aqueous glycerol solutions and

particle–water mixtures.

Although aqueous glycerol solutions are excellent

analogues for sediment–water mixtures they do not

reproduce all aspects that are important to flow be-

haviour. Aqueous glycerol solutions do not reproduce

the influence of non-Newtonian rheology on flow

behaviour. In particular, sedimentQwater mixtures

with relatively high concentrations of non-cohesive

particles or significant mud content may possess a

yield strength and display shear-thickening or shear-

thinning behaviour (Barnes, 1989; Coussot, 1997;

Major and Pierson, 1992). Of course, the settling of

sediment particles and erosion of the substrate is not

accounted in experiments using solutions.

3.2. Flow measurements

The flow velocity and concentration were mea-

sured at different heights within the flow at a position

2.5 m downstream of the inlet, and the current was

filmed at this location. Flow velocity and concentra-

tion measuring apparatus were held within a machined

holder that ensured individual probes were set parallel

both to the bed and to the tank walls.

3.2.1. Flow velocity

The downstream component of flow velocity was

measured using ultrasonic Doppler velocity profiling

(UDVP). This method derives velocity using the

Doppler shift in ultrasound frequency recorded from

small particles passing through the measurement vol-

ume (Takeda, 1991; Best et al., 2001). The velocity of

a particle is given by:

U ¼ cfD=2f0; ð1Þ

where c is the speed of sound in the fluid being

investigated, fD is the Doppler shift and f0 is the

ultrasound frequency. Ultrasonic probes simultaneous-

ly measure the velocity in 128 measurement volumes

positioned along the length of the ultrasound beam.

Flow velocity was recorded upstream of the probes,

thus instrumentation placed in the flow did not affect

measurements. A vertical array of eight probes was

used to measure the downstream velocity at heights of

0.1, 1.9, 2.8, 3.7, 5.7, 7.1, 8.5 and 11.4 cm above the

bed. The maximum temporal resolution of the velocity

data was 5.8 Hz; other parameters are listed in Table 1.

Velocity data was post-processed to account for the

effect of flow concentration on velocity measure-

ments; measurements were affected by variations in

glycerol concentration since the speed of ultrasound is

a function of concentration. In order to correct data the

sound velocity of glycerol solutions was empirically

derived using an ultrasonic thickness gauge. The

depth was initially measured in water and then again

in a solution. The velocity of sound in the solution,

csol, was calculated from

csol ¼xsol

xwat

�cwat;

�ð2Þ

Table 1

Starting parameters of the ultrasonic Doppler velocity profilers

Number of probes 8

Height above floor, cm 0.1, 1.9, 2.8, 3.7, 5.7, 7.1,

8.5, 11.4

Ultrasound frequency, MHz 2

Transducer and probe

diameter, mm

Two probe sizes used: 5, 8

and 10, 13

Measurement window, mm 12.4–107.1

Measurement bin length, mm 0.74

Velocity resolution, mm s�1 2.3–3.4

Height of nearest measurement

bin, mm

6.1 and 11.1

Height of furthest measurement

bin, mm

14.35 and 19.4

Ultrasound velocity, m s�1 1480

Sampling frequency/probe, Hz 5.8

The term bbinQ refers to a volume in which velocity measurements

are recorded.


where cwat is the sound of velocity in water and xwatand xsol is the depth measured in water and the solu-

tion, respectively. The velocity of sound of solutions

was found to be proportion to glycerol concentration,

C, so that csol=0.0034C +0.99. The results show that

changes in flow concentration had a relatively small

influence (b5%) on flow velocity measurements.

3.2.2. Flow concentration

Flow concentration was measured using an array of

five vertically-stacked siphons (internal diameter of

0.6 cm) to extract fluid samples at 0.5, 1.7, 3.0, 5.0

and 8.0 cm above the bed. Fluid samples were collect-

ed in beakers positioned on a table on a movable track.

During experiments the table was moved to collect

samples at 5-s intervals. The concentration of glycerol

of each sample was determined by measuring their

refractive index using a temperature controlled refrac-

tometer. The refractive index, RI, of glycerol solutions

is proportional to glycerol concentration, C, and at a

temperature of 20 8C, RI=0.0014C +1.3304.

The temporal record of concentration may become

distorted if the siphon flow rate changes during the

experiment, for example due to fluctuation in flow

velocity, density, and viscosity. This problem occurred

in experiments using relatively high concentrations,

z80% weight glycerol. In these experiments the dis-

charge per unit time from siphons measuring close to

the bed varied by up to five times. Considering changes

in the discharge rate, a cumulative offset of several tens

of seconds or more over the duration of the time of the

flow can be crudely estimated. However, since in these

flows an abrupt change in concentration could be clear-

ly identified based on colouration from video record-

ings, a temporal correction was made to calibrate the

measured concentration of the affected siphon. Smaller

fluctuations of siphon discharge were observed in other

experiments and at measuring positions higher above

the bed; however, these were not corrected for.

3.2.3. Reproducibility

The reproducibility of flow velocity and concen-

tration measurements was tested by repeating one

experiment using the same starting conditions (Fig.

5). The results indicate standard deviations of 10–25

mm s�1 for velocity measurements. These values are

calculated by taking a temporal mean over the time

period of quasi-steady flow, 0–25 s after the arrival of

the flow front. Differences in velocity are relatively

large in the probes positioned at 5.7 and 7.1 cm above

the bed and at times during the waning flow phase

later than 25 s. Trends recorded by the lowest three

UDVP probes are broadly similar for the first 20 s.

Concentration measurements on average have stan-

dard deviations of b3% weight glycerol. A somewhat

larger deviation is recorded by the lowest siphon

during the initial 10 s (Fig. 5A).

3.3. Experimental runs

A set of experiments were run with the initial start-

ing conditions shown in Table 2. The initial basal-layer

concentration (CL) was increased from ~20% to 90%

glycerol whilst the initial upper-layer concentration

(CU) was kept at a low value between 6.5% and

12.3%. Values of glycerol concentration can be com-

pared to the volume fraction, t, of sediment–water

mixtures based on a comparison of values of kinematic

viscosity. On this basis solutions of 10% glycerol are

equivalent to volume fraction of cohesionless particles

of ~0.2t and 20% and 90% glycerol is equivalent to

volume fractions of ~0.3t and ~0.6t, respectively.

3.3.1. Starting stratification

The strength of the stratification of two-layer flows

may be assessed using dimensionless ratios whose

value if small indicate a strong stratification whilst a

value of unity indicates a homogeneous mixture. The

0

20

40

60

C,a

gs%

0.5 1.7 3 5 8

0

100

200

300

u,m

m/s

Run 5 Run 5a

0

100

200

300

u,m

m/s

-5 0 5 10 15 20 25 30 35 40

t, s

-50

50

150

250

u,m

m/s

0

100

200

300

u,m

m/s

0

100

200

300

u,m

m/s

0

100

200

300

u,m

m/s

z = 1 cmstd = 23.0

z = 1.9 cmstd = 11.3

z = 2.8 cmstd = 13.2

z = 3.7 cmstd = 16.6

z = 5.7 cmstd = 17.7

z = 7.1 cmstd = 25.0

z, cm =

A.

B.

Fig. 5. Concentration (A) and velocity (B) data for multiple runs using the same starting conditions (experiments 5 and 5a). The lower and upper

layer had initial glycerol concentrations of ~60% and ~10%, respectively (Table 2). Data show that the reproducibility of data is within F3%

weight glycerol and velocity data is within F30 mm s�1. Standard deviations were calculated for individual measurement probes using the

average standard deviation of measurements taken between 0 and 25 s, during steady input of fluid into the tank. The abbreviations bagsQ andbstdQ are aqueous glycerol solution and standard deviation in mm s�1, respectively.


dimensionless density ratio between layers was de-

fined by Gladstone et al. (2004):

q4 ¼ qU � qa

qL � qa

¼ gUV

gLV; ð3Þ

where qL and qU are the densities of the lower and

upper layers, respectively, qa is the density of the

ambient fluid and the reduced gravity gV=g(qf� qa /

qa) where g is the acceleration due to gravity, qf is the

density of the flow and qa the density of the ambient

fluid. We introduce a dimensionless viscosity ratio

defined as

l4 ¼ lU � la

lL � la

; ð4Þ

where lL and lU are the viscosities of the lower and

upper layers, respectively, and la is the viscosity of the

ambient fluid. Gladstone et al. (2004) also defined a

Table 2

Starting parameters of the experiments

Experiment CL CU qL qU q* lL lU l* QL QU B* Ta TL TU

1 18.3 9.8 1033.6 1010.9 0.50 0.0020 0.0014 0.446 0.0057 0.005 0.30 11.2 14.4 14.0

2 28.3 10.2 1060.2 1011.9 0.33 0.0026 0.0015 0.331 0.0057 0.005 0.22 10.4 16.6 14.2

3 37.2 6.9 1084.0 1003.1 0.16 0.0038 0.0013 0.116 0.0054 0.005 0.12 11.2 17.6 14.7

4 44.7 8.5 1104.0 1007.4 0.17 0.0054 0.0014 0.096 0.0047 0.005 0.15 10.1 19.3 14.9

5 58.5 9.8 1140.8 1010.9 0.15 0.0120 0.0014 0.038 0.0053 0.005 0.12 10.9 17.6 16.2

5a 59.5 9.4 1143.5 1009.8 0.14 0.0130 0.0014 0.033 0.0054 0.005 0.11 10.4 20.5 14.2

6 67.9 8.5 1165.9 1007.4 0.11 0.0207 0.0014 0.021 0.0054 0.005 0.09 11.6 20.2 15.9

7 79.4 12.3 1196.5 1017.6 0.14 0.0624 0.0015 0.009 0.0030 0.005 0.19 10.5 21.5 15.7

8 87.3 6.5 1217.6 1002.1 0.06 0.1402 0.0013 0.002 0.0020 0.005 0.13 10.7 19.6 15.5

9 88.3 7.7 1220.3 1005.3 0.07 0.1350 0.0014 0.003 0.002 0.005 0.16 11.1 21.1 15.1

The variables are glycerol concentration, C, in percentage by weight; density, q, in kg m�3; dimensionless density ratio, q*; viscosity, l, in kg

m�1 s�1; dimensionless viscosity ratio, l*; discharge, Q, in m3; and dimensionless buoyancy ratio, B*. Subscripts a, L and U indicate values

for the ambient fluid and lower and upper layers of the current, respectively. Viscosity values are corrected for temperature using the empirical

relationship found by Chen and Pearlstein (1987).


ratio for the difference in the driving buoyancy of each

layer, B*, proportional to the layer density and volume.

For continuous input flows B* can be defined as

B4 ¼ QUgUV

ðQUgUVþ QLgLVÞ; ð5Þ

where Q is the discharge per unit width. Values of

0bB*b0.5 indicate a greater driving buoyancy in the

lower layer, whilst those of 1NB*N0.5 indicate a

greater driving buoyancy in the upper layer. The start-

ing conditions were chosen to explore a distinct area of

the bparameter spaceQ of stratified flows where q*, l*and B* are b0.5 (Table 2). The results from these

experiments thus document currents with density

ratios of 0.06bq*b0.50 and viscosity ratios of

0.002bl*b0.48 and those with a greater driving buoy-ancy in their lower layer, 0bB*b0.3. In order to keep

input conditions similar, the upper layer was released

slightly, up to 5 s, before the lower layer in experiments

1–8. In two experiments, 9 and 10, the lower layer was

released first in order to see how this affected flow

behaviour (Table 2).

4. Experimental results

4.1. Visual observations

Gravity currents with a characteristic head and

body structure were formed after releasing the glyc-

erol solutions. The two constituent layers were ob-

served to have variable downstream velocities. Thus,

two distinct types of current developed with either a

faster lower layer or a faster upper layer. In each case

the faster layer ran ahead to form the flow front. In

some experiments, overtaking of one layer by the

other was observed. This occurred within 1 m of the

input point and upstream of where flow measurements

were recorded. The basal layer, when faster, overtook

by pushing lighter fluid of the slower layer upwards

and out of the way (Fig. 6A). The upper layer, when

faster, overtook by propagating over the relatively

slow-moving lower layer (Fig. 6D). The upper layer

became progressively thinner and increasingly

stretched-out as it moved. On approaching the flow

front it intruded into the back of the head along a

density interface between the denser fluid of the

lower layer and the lighter fluid of the wake. In other

experiments overtaking was not observed since the

faster layer was released first (Fig. 6B and C). After

the flow front had reached the end of the tank floor, a

longitudinally uniform current developed, whose two

layer stratification was preserved along the length of

the tank.

4.1.1. Flows with a fast lower layer

A current with a relatively fast basal layer was

formed in experiments run with lower-layer concentra-

tions less than 75% glycerol (experiments 1–6). In

these flows fluid from the lower layer formed the

head of the current (Fig. 7A). Video recordings show

A. Upper layer released first but overtakenby faster lower layer. Experiments 1-6.

B. Upper layer released first and remainingahead of slower lower layer. Experiment 7 & 8.

C. Lower layer released first and remainingahead of slower upper layer. Experiment 10.

D. Lower layer released first but overtakenby faster upper layer. Experiment 9.

Basal fluid Top fluid Mixed fluid

Fig. 6. Schematic diagram showing the evolution of two-layer, stratified gravity currents based on experimental observations. Four types of

evolution were observed depending on which layer had a faster velocity and which was released first.


that only a small amount of fluid from the upper layer

was able to intrude into the head, instead most of this

fluid was swept back into the wake before reaching the

flow front. Behind the head, the lower layer had an

average thickness of 2–3 cm. The upper layer was

thicker, on average being between 6 and 8 cm thick.

The interface varied in character between experiments

becoming progressively distinct and sharply defined

with increasing lower-layer concentration. In all

experiments interfacial waves developed and thus the

height of the interface varied temporally. In experi-

ments 6 and 7 wave heights were of a similar scale to

the thickness of the lower layer. Tracer particles within

all flows moved in a turbulent fashion, changing both

height above the bed and speed. Mixing was clearly

visible between layers. In relatively high-concentra-

tion flows (e.g., experiments 5 and 6) mixing occurred

in periodic bursts whereby fluid from the denser lower

layer was injected upwards into the layer above.

4.1.2. Flows with a fast upper layer

The lower layer was relatively slow compared to

the upper layer in experiments run with lower-layer

concentrations greater than 75% glycerol (experi-

ments 7 and 8). In these flows the lower layer formed

a slow-moving region that lagged behind the current’s

head. The current’s head was formed by fluid fed from

the upper layer (Fig. 7B). In experiments 7 and 8 it

took the lower layer 8 and 11 s to arrive at the

measurement station after the passage of the head,

respectively. This slow-moving region had a wedge-

shaped front inclined downstream (Fig. 7B). With

time, the lower layer achieved a constant thickness

of several centimetres. In experiment 7 the interface

was noticeably wavy whilst in experiment 8, with the

highest glycerol concentration (CL=90%), the inter-

face between the two layers was flat. Many of the

tracer particles were observed to become concentrated

at the interfacial boundary. For experiment 8, those

Fig. 7. Successive photos taken at a position of 2.5 m downstream of the inlet point. (A) A current with a relatively fast lower layer (experiment 3). Photographs were taken at 2-s

intervals. (B) A current with a relatively fast upper layer and a relatively slow, laminar-moving lower layer that lags behind the flow front (experiment 8). Photographs were taken

every 4 s.

L.A.Amyet

al./Sedimentary

Geology179(2005)5–29

14


tracer particles on the boundary and within the lower

layer were observed to move in a laminar fashion

whereby they maintained a constant speed and

moved in a straight line at a constant height above

the bed. Mixing between the two layers, especially in

experiment 8, appeared to be suppressed. However,

fluid of a colour indicative of mixing and moving

relatively fast was observed preceding the arrival of

the lower layer.

-10 0 10 20

t, s

Run 8, lower layer ~ 90

02468

10

z, c

m

-10 0 10 20

Run 3, lower layer ~ 40 %

02468

10

z, c

m

-10 0 10 20


02468

10

z, c

m

-10 0 10 20


02468

10

z, c

m

-10 0 10 20


02468

10

z, c

m

-10 0 10 20


02468

10

z, c

m

A.

Fig. 8. (A) Maps of flow velocity constructed from temporal measurement

downstream of the inlet point for selected experiments. Data shown is the

flow concentration constructed from temporal measurements taken at five h

inlet point for selected experiments. The time, t, is measured in seconds afte

conditions for limited periods of time some 5 to 20 s after the arrival of t

concentration lower layer. Experiments 7 and 8 had a relatively slow-mov

4.2. Velocity and concentration profiles

Maps of flow velocity and concentration through

time and for different heights above the bed are shown

in Fig. 8. These plots show that currents achieved

quasi-steady conditions for both flow velocity and

concentration for a period of 10 s or more. The initial

recorded flow velocity and concentration, however,

are unsteady and typically waxing for the first 5 to

30 40 50 60

% glycerol

0

40

80

120

160

200

240

280

30 40 50 60

glycerol

0

40

80

120

160

200

240

280

30 40 50 60

glycerol

0

40

80

120

160

200

240

280

30 40 50 60

glycerol

0

40

80

120

160

200

240

280

30 40 50 60

glycerol

0

40

80

120

160

200

240

280

mm/s

30 40 50 60

glycerol

0

40

80

120

160

200

240

280

s taken at eight heights (z) above the bed and at a position of 2.5 m

mean velocity of 60 bins and three successive cycles. (B) Maps of

eights (z) above the bed and at a position of 2.5 m downstream of the

r the arrival of the flow front. Measurements show quasi-steady flow

he flow front. Experiments 1, 3, 5 and 6 had a relatively fast, high-

ing, high-concentration, lower layer.

-10 0 10 20 30 40 50 60

Run 1, lower layer ~ 20 % glycerol

02468

10

z, c

m

0.0

4.0

8.0

12.0

16.0

20.0

- 10 0 10 20 30 40 50 60


02468

10

z, c

m

0.0

12.0

24.0

36.0

48.0

60.0

-10 0 10 20 30 40 50 60


02468

10

z, c

m

0.0

16.0

32.0

48.0

64.0

80.0

-10 0 10 20 30 40 50 60


02468

10

z, c

m

0.0

8.0

16.0

24.0

32.0

40.0

-10 0 10 20 30 40 50 60

t, s


02468

10

z, c

m

-10.0

10.0

30.0

50.0

70.0

90.0

Conc. % ags

-10 0 10 20 30 40 50 60


02468

10

z, c

m

0.0

14.0

28.0

42.0

56.0

70.0

B.

Fig. 8 (continued).


15 s (Fig. 8). This is most apparent in the lower part of

the flow immediately above the bed. Flow unsteadi-

ness at these times is related to the passage of the

current’s head and subsequent large-scale eddies. Cur-

rents began to wane after about 20 s marking the time

at which the input supply was turned off (Fig. 8).

The waning of flows becomes noticeably stronger

with increasing lower-layer concentration.

4.2.1. Flows with a fast lower layer

Experimental flows 1–6 with fast moving bases are

characterised by temporal concentration and velocity

profiles that mirror one another in that maximum and

minimum values occur at similar times (Fig. 8). Near-

bed velocities and concentrations tend to increase with

time to a quasi-steady state before waning (Fig. 9).

Maximum measured velocities of ~300 mm s�1 oc-

curred in experiments 6 and 7 with lower-layer gly-

cerol concentrations of 60% and 70%, respectively.

Vertical profiles in the body (Fig. 10) are similar to

those described for other low-concentration currents

(Fig. 2A–B); concentration displays continuous or

nearly continuous profiles whilst velocity has a con-

cave upward-facing shape above the velocity maxi-

mum (Fig. 10A–D). The velocity maximum occurs

between 1 to 2 cm above the bed and at a fraction of

0.1–0.15 of the total flow depth. The velocity maxi-

mum in experiments 1 to 6 occurs at a similar height

to the interfacial boundary between layers. Root-

mean-squared (RMS) values of the temporal velocity

-10 0 10 20 30 40 50 60

t, s

0

100

200

300

020406080100

020406080100

C, %

ags

020406080100

0

100

200

300

020406080100

C, %

ags

020406080100

0

100

200

300

020406080100

C, %

ags

020406080100

0

100

200

300

020406080100

C, %

ags

020406080100

0

100

200

300

020406080100

C, %

ags

020406080100

C, %

ags

0

100

200

300

B. Run 3 (40%)

A. Run 1 (20%)

C. Run 5 (60%)

D. Run 6 (70%)

E. Run 7 (80%)

F. Run 8 (90%)

u, m

m/s

u, m

m/s

u, m

m/s

u, m

m/s

u, m

m/s

u, m

m/s

Fig. 9. (A–F) Temporal profiles of velocity and concentration measured at 1 cm and 0.5 cm above the bed, respectively, and at a position of 2.5

m downstream of the inlet point, for selected experiments. The time, t, is measured in seconds after the arrival of the flow front. The dashed line

of flow concentration for experiment 8 shows data that have been corrected for temporal displacement resulting from variable siphon flow rates.

See text for further explanation.


time-series, a proxy for flow turbulence, have maxi-

mum values close to the bed and show a decrease

upwards away from the bed. This type of distribution

has been observed in other quasi-steady turbulent

density currents (e.g., Buckee et al., 2001).

4.2.2. Flows with a fast upper layer

The temporal and vertical profiles of experimen-

tal currents 7 and 8 with slower moving lower

layers are markedly different to those with faster

ones. In these currents the flow velocity and con-

0 100 200 300

u, mm/s

0

2

4

6

8

10

12

z,cm

0 20 40 60 80 100

C, % ags

0 100 200 300

u, mm/s

0 10 20 30 40 50

Urms, mm/s

u, mm/s u, mm/s C, % ags Urms, mm/s

0

2

4

6

8

10

12

0 100 200 300

u, mm/s

0 20 40 60 80 100

C, % ags

0 10 20 30 40 50

Urms, mm/s

0

2

4

6

8

10

12

0 100 200 300

u, mm/s

0 20 40 60 80 100

C, % ags

0 10 20 30 40 500 10 20 30 40 50

Urms, mm/s

0

2

4

6

8

10

12

0 100 200 300

u, mm/s

0 20 40 60 80 100

C, % ags

0 10 20 30 40 50

Urms, mm/s

0

2

4

6

8

10

12

0 100 200 300

u, mm/s

0 20 40 60 80 100

C, % ags

0 10 20 30 40 50

Urms, mm/s

B. Run 3 (40%) E. Run 7 (80%) F. Run 8 (90%)C. Run 5 (60%)

0 20 40 60 80 1000

2

4

6

8

10

12

z,cm

0

2

4

6

8

10

12

0 20 40 60 80 1000 20 40 60 80 1000 20 40 60 80 1000

2

4

6

8

10

12

0 20 40 60 80 1000 20 40 60 80 1000 20 40 60 80 1000

2

4

6

8

10

12

0 20 40 60 80 1000 20 40 60 80 1000 20 40 60 80 1000

2

4

6

8

10

12

0 20 40 60 80 1000 20 40 60 80 1000 20 40 60 80 100

G. Run 1 (20%) H. Run 3 (40%) K. Run 7 (80%) L. Run 8 (90%)I. Run 5 (60%)

0

2

4

6

8

10

12

0 100 200 300

u, mm/s

0 20 40 60 80 100

C, % ags

0 10 20 30 40 50

Urms, mm/s

D. Run 6 (70%)

0

2

4

6

8

10

12

0 20 40 60 80 1000 20 40 60 80 1000 20 40 60 80 100

J. Run 6 (70%)

Dimensional values

Normalised values

A. Run 1 (20%)

υ, σ , χ υ, σ , χ υ, σ , χ υ, σ , χ υ, σ , χ υ, σ , χ

υ συ χ

L.A.Amyet

al./Sedimentary

Geology179(2005)5–29

18


centration do not mimic one another (Fig. 8). Near

bed velocities increase with time and then decrease

whilst concentration is initially low before increas-

ing (Fig. 9). In experiment 8 an abrupt increase in

flow concentration is seen at about 15 s, marking

the arrival of the slow-moving lower layer (Fig.

9F). An area of enhanced flow velocity also occurs

several centimetres above the bed before the arrival

of the slow-moving lower layer at 7–12 s (Fig. 8).

Initial flow unsteadiness is related to the passage of

the head but also to their longitudinal structure.

Vertical concentration profiles of experiments 7

and 8 display a less gradual and more step-like

stratification than experiments with lower concen-

trations, e.g. experiments 1 and 3 (Fig. 10E–F).

Currents with a slow moving lower layer have a

convex shaped velocity distribution above the max-

imum being quite different to those recorded for

experiments 1–6. Also in experiment 8, the velocity

maximum is relatively high in the flow and it is

situated in the upper-layer at a fractional depth of

between 0.3 and 0.4 (Fig. 10L). This velocity

distribution is similar to those of high-concentration

suspension flows reported by Postma et al. (1988).

RMS values of the temporal velocity time-series

show profiles and values similar to currents with

a faster lower layer; maximum values occur close

to the lower flow boundary and values decrease

upwards. The high RMS values recorded near the

bed are surprising, especially for experiment 8,

since visual observations suggest that the lower

layer moved in a laminar fashion. Since the veloc-

ity measurements taken by the lowest position

probe were taken from an area close to the inter-

facial boundary between layers, we suggest that the

high RMS values are caused by turbulence related

to the interfacial boundary. Alternatively, conditions

close to the tank wall may have been different to

those in the centre of the tank where the data were

recorded; the interfacial boundary may have been

lower or the lower layer more turbulent in the

centre of the tank.

Fig. 10. (A–F) Vertical profiles of the downstream velocity (u), the root-m

for selected experiments. Profiles are taken from the body of currents at 10

was taken at 20 s. Velocities indicated by crosses show values corrected

uncorrected values. (G–L) Non-dimensional values given as a percentage o

mean-square of downstream velocity (r) and concentration (v).

5. Discussion

The experiments show that the behaviour of con-

tinuously-fed gravity currents is strongly controlled

by their stratification. Initial flow unsteadiness is re-

lated to the passage of the head, but also to the

current’s longitudinal structure. In the ranges of q*,l* and B* (all b0.5) investigated, two distinct types

of behaviour were observed. A summary of the char-

acteristics of each current type is shown in Fig. 11.

Those with low to moderate maximum concentra-

tions, b75% glycerol, have fast-moving, high-concen-

tration, basal regions (Fig. 11A). In contrast, currents

with relatively high maximum concentrations, N75%

glycerol, have a slow moving basal region that lags

behind the flow front (Fig. 11B). Particle-laden labo-

ratory currents with similar flow structures to these

two types have been observed previously. Flows with

high-concentration fast-moving bases have been noted

by Hampton (1972), Mohrig et al. (1998) and Marr et

al. (2001), whilst currents with slow moving bases

have been observed by Postma et al. (1988).

5.1. Interpretation of flow behaviour

We interpret the observed change in behaviour to

correspond to a transition in the flow dynamics of the

lower layer from being inertia-driven to viscosity-

controlled. A reduction in the lower layer velocity at

concentrations exceeding 75% glycerol can be ex-

plained by the enhanced drag at the lower flow bound-

ary, a characteristic of high viscosity flows. The dimen-

sionless Reynolds number (Re) is a measure of the ratio

of inertial to viscous forces and may be used to assess

the transition between turbulent and laminar flow re-

gimes. The Reynolds number is evaluated here using

Re ¼ uhql

ð6Þ

where u is a velocity scale, h is a length scale and qand l are the density and viscosity of the fluid,

respectively. The drag force experienced by flows

ean-square of downstream velocity (Urms) and the concentration (C)

s after the arrival of the flow front except the profile for Run 8 which

for fluid density (see text) whilst those indicated by squares show

f the maximum value in each profile; downstream velocity (t), root-

Time

u,c

Velocity

Concentration

iii

Time

u,c

Velocity

Concentration

A. Current with a relatively fast-moving high-concentration phaseE.g., experiments 1-6

X Z

B. Current with a relatively slow-moving high-concentrationE.g. Experiments 7 and 8

Vertical bed structure

Vertical bed structure

High-concentration Low-concentration

Y

Fig. 11. Summary diagram of experimental data showing the two different current types observed. (A) A current with a relatively fast-moving,

high-concentration lower layer that also forms the flow front. (B) A current with a relatively fast-moving, high-concentration upper layer. The

lower-layer lags behind the flow front. For each current type the near-bed temporal trends of velocity and concentration and corresponding

inferred depositional sequence at a single location is shown.



varies with Reynolds number. The amount of drag is

similar in turbulent flows with relatively high Rey-

nolds numbers but increases significantly at smaller

Reynolds numbers in the range of transitional and

laminar flow conditions. This relationship between

drag and Reynolds number is known to apply to

pipe flow (Chadwick and Morfett, 1992). van Kessel

and Kranenburg (1996) showed that the drag coeffi-

cient varies strongly at low Reynolds numbers,

Re b103, for gravity currents of fluidQmud (Fig. 12).

The Reynolds number of gravity currents is usually

deduced using the bulk current properties to yield a

single value for the whole current. Since viscosity-

stratified currents may exhibit both laminar and tur-

bulent flow at different heights above the bed, the

application of a single Reynolds number may not be

appropriate. One way to assess the flow character of

strongly stratified currents is to calculate separate

Reynolds numbers for regions near the bed and higher

up in the flow (Table 3A). Reynolds numbers were

calculated for below (RebUmax) and above (ReNUmax

)

the velocity maximum using measured flow properties

Re ¼ UhP

M; ð7Þ

where U, P, and M are the layer depth-average veloc-

ity, concentration, and viscosity respectively, and h is

10 100 1000 10000 100000

Ree

0.001

0.01

0.1

1

CD

Fig. 12. Drag coefficient, CD, as a function of the effective Rey-

nolds number, Ree, for fluid mud gravity currents presented by van

Kessel and Kranenburg (1996). Their original data is approximated

by the curve CD~(12+0.1Ree)/Ree, where Ree is an effective

Reynolds number (see Eq. (11) in van Kessel and Kranenburg,

1996). The graph shows that drag increases significantly at low

Reynolds numbers and in the range below turbulent flow conditions,

Reeb3000, using the criterion proposed by Liu and Mei (1990).

the layer depth (Table 3A). Various values of the

Reynolds number have been proposed for the thresh-

old between laminar to turbulent flow ranging over

one order of magnitude. However, laboratory studies

often take the transitional number as 500–2000 (Simp-

son and Britter, 1979; Allen, 1985). This threshold is

also used in this study and given the range of values

calculated for the experimental currents, would appear

to be a reasonable approximation. Values above the

velocity maximum fall into the turbulent regime,

3000bRe b12000, corresponding to observations of

turbulent particle movement in the upper part of the

flow. Smaller values of the Reynolds number charac-

terise the region below the velocity maximum. Impor-

tantly, they show a decrease from 2500 to 900 Re with

increasing glycerol concentration. These Reynolds

numbers fall into the range where values of drag are

expected to vary strongly and support our interpreta-

tion of a varying drag influence at the lower flow

boundary. The flow Reynolds numbers calculated are

subject to some measurement error in flow concentra-

tion. The value of the Reynolds number below the

velocity maximum for experiment 7 (with the second

highest glycerol concentration) is spuriously high at

1600 compared to values for flows of similar concen-

tration. This corresponds to relatively low values of

glycerol concentration recorded for this flow and is

likely to be an artefact of selective siphoning of fluid

with a lower concentration and viscosity. Reynolds

numbers calculated using the initial values of fluid

density and viscosity for the lower layer (ReL), may

therefore give a better estimate for high-concentration

flows in which mixing was unimportant (Table 3B).

These values indicate Reynolds numbers for the high-

concentration flows of b100, consistent with visual

observations of laminar particle movement in these

currents.

5.2. Stratified flow regimes

Based on experimental data, Gladstone et al.

(2004) constructed a flow regime diagram for the

behaviour of two layer, stratified currents of various

density ratios, q* and buoyancy ratios, B* (Fig. 13A).

In their scheme, the layer buoyancy determines which

layer runs ahead to form the leading edge of the flow;

the lower and upper layers run ahead when B*b0.5

and B*N0.5, respectively. All the currents studied

Table 3

Calculated flow characteristics for laboratory experiments

A. Parameters based on depth-averaged values and Umax subdivision

Experiment UC hC PC MC ReC Fr RiB UbUmaxhbUmax

PbUmaxMbUmax

RebUmaxUNUmax

hNUmaxPNUmax

MNUmaxReNUmax

1 0.10 0.13 994.1 0.0011 11,768 1.1 0.9 0.154 0.02 1004.9 0.0012 2513 0.08 0.11 992.2 0.0011 7768

2 0.10 0.13 1000.6 0.0012 10,510 0.7 1.3 0.136 0.02 1017.5 0.0014 2034 0.09 0.11 997.5 0.0011 8486

3 0.11 0.13 997.4 0.0011 12,069 1.0 1.0 0.116 0.02 1025.5 0.0015 1553 0.09 0.11 992.3 0.0011 9000

4 0.10 0.13 995.1 0.0011 11,598 1.0 1.0 0.130 0.02 1027.3 0.0016 1635 0.08 0.11 989.3 0.0011 8360

5 0.11 0.13 1003.5 0.0012 11,732 0.8 1.3 0.135 0.02 1025.5 0.0015 1800 0.09 0.11 999.5 0.0012 8388

6 0.10 0.13 1003.6 0.0012 10,990 0.7 1.4 0.105 0.02 1064.2 0.0026 847 0.09 0.11 992.6 0.0011 8873

7 0.13 0.13 1001.9 0.0012 14,539 1.0 1.0 0.135 0.02 1044.9 0.0017 1623 0.11 0.11 994.1 0.0011 11,266

8 0.07 0.10 1011.0 0.0013 5522 0.5 2.1 0.095 0.03 1094.1 0.0037 855 0.06 0.07 991.1 0.0011 3664

B. Parameters based on initial starting values and layer subdivision

Experiment UC hC qC lC ReC Fr RiB UbUmaxhL qL lL ReL UNUmax

hU qU lU ReU

1 0.10 0.13 1014.4 0.0015 8804 0.5 1.8 0.154 0.02 1033.6 0.0020 1622 0.08 0.11 1010.9 0.0014 6012

2 0.10 0.13 1019.4 0.0017 7470 0.5 2.1 0.136 0.02 1060.2 0.0026 1114 0.09 0.11 1011.9 0.0015 6454

3 0.11 0.13 1015.6 0.0017 8305 0.6 1.8 0.116 0.02 1084.0 0.0038 669 0.09 0.11 1003.1 0.0013 7497

4 0.10 0.13 1022.3 0.0020 6559 0.5 2.1 0.130 0.02 1104.0 0.0054 531 0.08 0.11 1007.4 0.0014 6303

5 0.11 0.13 1030.9 0.0031 4796 0.5 2.2 0.135 0.02 1140.8 0.0120 255 0.09 0.11 1010.9 0.0014 6994

6 0.10 0.13 1031.8 0.0044 3129 0.4 2.3 0.097 0.02 1165.9 0.0207 109 0.09 0.11 1007.4 0.0014 6925

7 0.13 0.13 1045.1 0.0109 1666 0.5 2.0 0.135 0.02 1196.5 0.0624 52 0.11 0.11 1017.6 0.0015 8368

8 0.07 0.10 1066.7 0.0430 176 0.3 3.9 0.095 0.03 1217.6 0.1402 25 0.06 0.07 1002.1 0.0013 3022

(A) Flow characteristics calculated using depth-averaged values for the whole current and for portions of the current above and below the

velocity maximum. Variables are: depth-averaged velocity, U, in m s�1; height, h, in m; depth-averaged density, P, in kg m�3; viscosity, M, in

kg m�1 s�1; dimensionless Reynolds number, Re =UCPh/M; dimensionless Froude number, Fr =UC/(hC(gP�q0/q0)1/2 (where g is the

acceleration due to gravity and q0 is the density of the ambient fluid); and dimensionless Richardson number, RiB=1/Fr. Subscripts C, NUmax

and bUmax indicate values for the whole current, and the portion of the current below and above the velocity maximum, respectively. Current

height, hC, was estimated from photos. The height of the velocity maximum was estimated using velocity measurements (Fig. 10). (B) Flow

characteristics calculated using initial values for the whole current and for upper and lower layers. Variables are: depth-averaged velocity, U, in

m s�1; height, h, in m; layer density, q, in kg m�3; viscosity, l, in kg m�1 s�1; dimensionless Reynolds number, Re =UCqh/l; dimensionless

Froude number, Fr =UC/(hC(gqC�q0/q0)1/2; and dimensionless Richardson number, RiB=1/Fr. Subscripts L, U and C indicate values for

lower and upper layers and whole current, respectively. Note that the velocity values used are the same in both A and B.


here had a greater buoyancy in their lower layer with

values of B*b0.5. Experiments 1 to 6 with a relative-

ly fast-moving lower layer displayed behaviour con-

sistent with the findings of Gladstone et al. (2004) for

inertial, lock-release currents. In experiments 7 and 8,

the lower layer was relatively slow-moving compared

to the upper layer, despite it having a much greater

driving force, B*b0.2. This indicates that the flow

buoyancy does not control flow behaviour in currents

with strong viscosity stratification l* (of the order of

1�10�3 in the present experiments). Hence, a third

axis to Gladstone et al.’s proposed regime diagram

may be added to describe currents with varied viscos-

ity stratification (Fig. 13B).

The second axis of the regime diagram constructed

by Gladstone et al. (2004) indicates relative amounts

of mixing between layers based on q*; for currents

with q*b0.4 the initial stratification is maintained

whilst for those with q*N0.4 stratification is quickly

destroyed by mixing. Comparison between the present

experiments and those of Gladstone et al. (2004), in

terms of q*, however, is not straightforward for sev-

eral reasons. Firstly, the degree of mixing in continu-

ous-flux flows will be significantly different to surge-

type flows since the latter is dominated by the

dynamics of the current’s head. Secondly, the entire

duration of flow in Gladstone et al.’s study was ob-

served from the point of initiation to their arrest where-

as in the present experiments, currents could not be

observed to their natural stopping point. Bearing this in

mind, it was observed that in all flows with density

ratios b0.35, the initial two layer stratification was

maintained, this being consistent with Gladstone et

al. (2004) results.

Fig. 13. (A) Regime diagram constructed by Gladstone et al. (2004) summarising the behaviour of two-layer density-stratified, surge-type

currents. The graph describes currents in terms of the dimensionless parameters density ratio q* and distribution of buoyancy B*. Modified

from Gladstone et al. (2004). (B) Diagram showing the parameter space varied in the present set of experiments in terms of the three

dimensionless parameters, density ratio q*, distribution of buoyancy B* and viscosity ratio l*. The range in these parameters explored is

shaded. Schematic cartoons for current behaviour are shown for different regimes of flow behaviour.



5.3. Implications for deposit character

The vertical characteristics of sediment gravity

flow deposits are controlled by temporal variations

in flow velocity and concentration at a point, and

thus, by the stream-wise structure of the current

(Branney and Kokelaar, 2002; Kneller and McCaf-

frey, 2003; McCaffrey et al., 2003; Choux et al.,

2005—this issue). The experimental data show that

stratified flows with relatively fast and slow-moving

lower layers have markedly different near-bed tem-

poral trends in flow properties (Figs. 9 and 10).

Assuming natural sediment-laden flows also display

these current structures, it follows that several dis-

tinct bed types should be deposited. Here we spec-

ulate on the characteristics of these beds, and

propose two simple depositional models for stratified

sediment-laden currents (Fig. 11). These models as-

sume deposition occurs throughout the passage of

the head, body and tail of the flow and at a single

location. The models developed should be consid-

ered as didealT deposit types. As for other models,

such as the Bouma sequence, variations should be

expected given the range of controlling factors on

the final deposit character.

5.3.1. Deposit interpretation

In order to identify the deposits of stratified

submarine currents, sedimentary features that reli-

ably record deposition by flows of low and high

sediment concentration need to be defined. This has

been a controversial subject, especially with regard

to the interpretation of massive sandstones (see dis-

cussions in Kneller and Buckee, 2000; Mulder and

Alexander, 2001; Amy et al., 2005—this issue).

Gradual deposition of particles from a relatively

low sediment-concentration phase (turbidity current)

may be recognised by the presence of tractional

bedforms, vertical normal grading under waning

flow conditions and a high degree of grain-size

sorting within the deposit. However, structureless

intervals such as the Bouma Ta division may be

produced under relatively high sediment-load fall-

out rates leading to the suppression of traction

(Arnott and Hand, 1989). In circumstances of steady

flow the Ta division may also lack grading (Kneller,

1995). Deposition from a high-concentration phase

(debris flow) occurring by en masse settling will

also tend to produce ungraded beds. However,

given a wide grain-size distribution (mud to centi-

metre clasts), debris flow deposits will be distinctive

from Bouma Ta divisions on account of their poor

sorting, relatively high matrix mud contents and

randomly distributed outsized clasts. In addition

debrites may display a distinct shear fabric. These

criteria follow those used by others (e.g., Lowe,

1982; Ghibaudo, 1992; Mulder and Alexander,

2001).

5.3.2. Deposits from flows with fast-moving high-

concentration phases

Particle-laden currents with relatively fast moving,

high-concentration phases will have an initial phase

of deposition from flow with relatively high sediment

concentrations followed later by deposition from

flow with lower sediment concentrations. The result-

ing depositional sequence will comprise a high-

concentration flow deposit overlain by a low-concen-

tration flow deposit (Fig. 11A). Beds displaying this

type of vertical character and ranging from several

metres to over tens of metres in thickness are com-

monly exposed in the Eocene/Oligocene sedimentary

sequence of the Gres de Peıra Cava Formation, SE

France (Fig. 14A). These sediments were deposited

in a relatively small (tens of kilometres long and

wide) deep-water basin in which flows were confined

by the local basin bathymetry (Hilton, 1994; Amy,

2000; McCaffrey and Kneller, 2001; Amy et al.,

2004). In relatively proximal sections, many beds

contain a coarse-grained (small-pebble to very coarse

sand grade), very poorly sorted, clast-rich basal in-

terval. The basal portion of these beds is interpreted

as having been deposited from a high sediment-con-

centration flow phase. The mud content varies later-

ally in the basal interval, implying that the cohesive

strength of the flow may have varied locally. The

upper part of the beds is finer grained and better

sorted and displays normal grading and usually cur-

rent lamination. The upper parts of beds are inter-

preted to record deposition from a relatively low-

concentration portion of the current deposited after

the high-concentration phase had passed. Correla-

tions indicate that these are the deposits of single

flow events and not an amalgamated sequence pro-

duced by multiple flows of different particle concen-

trations (Amy, 2000).


5.3.3. Deposits from flows with slow-moving high-

concentration phases

Particle-laden currents with relatively slow-mov-

ing, high-concentration phases will deposit initially

Fig. 14. Examples of beds preserved in ancient turbidite successions. (A)

France. This bed is interpreted as the deposit of a stratified sediment grav

Sandstone bed with a tripartite bed structure from the Marnoso-arenacea,

been deposited by a stratified sediment gravity flow with a slower-moving

from flow with low to intermediate sediment concen-

trations, followed by deposition from flow with high

sediment concentrations and finally from the trailing

flow with relatively low sediment concentrations. At a

Sandstone bed from the Gres de Peıra Cava, Maritime Alps of SE

ity flow with a faster-moving high-concentration lower region. (B)

northern Apennines of Italy. This type of bed is interpreted to have

, high-concentration, lower region. See text for further explanation.


single location the deposit will show a tripartite struc-

ture with the deposit of the high-concentration phase

encased between the deposits of the relatively low-

concentration flow phases (Fig. 11B). A significant

proportion of sediment gravity flow deposits of the

Miocene, Marnoso-arenacea Formation located in the

Italian Apennines display this type of tripartite bed

structure. These beds have been interpreted as debris

flow deposits sandwiched between turbidites (Ricci

Lucchi and Valmori, 1980; Talling et al., 2004; Amy

et al., 2005—this issue) formed in an open basin-plain

environment of the Apennine foredeep (Ricci Lucchi

and Valmori, 1980; Argnani and Ricci Lucchi, 2001).

The basal interval is composed of b20 cm, mud-poor

(b15% in thin section; Talling et al., 2004), coarse- to

fine-grained sandstone (Fig. 14B). The middle debrite

sandstone interval is usually slightly finer and thicker

(~20–90 cm) than the basal interval and relatively

mud-rich (15–22% in thin section; Talling et al.,

2004). It contains floating out-sized clasts several

millimetres to tens of centimetres in diameter. The

upper division is usually relatively thin (b20 cm),

fine- to very fine-grained sandstone with millimetre-

scale cross-lamination or parallel lamination.

Evidence that these deposits record a single flow

event rather than several amalgamated event beds are

(a) that the clast-rich debris flow units always occur

within this tripartite vertical bed sequence and (b)

long-distance correlations show that tripartite beds

do not dbreak-apartT into individual beds moving lat-

erally (Talling et al., 2004; Amy et al., 2005—this

issue). Correlations also show that the middle clast-

rich interval pinches out rapidly (over b5 km) down-

stream (Talling et al., 2004). This geometry suggests

that this portion of the bed was deposited en masse by

a high-concentration flow. In comparison, the lower

interval extends downstream of the pinch-out position

of the middle interval (Talling et al., 2004). Beds with

a similar tripartite vertical bed profile have been de-

scribed from the Pennsylvanian Jackfork Group in

Kansas (Hickson, 1999), Jurassic fans in the North

Sea (Haughton et al., 2003), and the Miocene and

lower Pliocene Laga Formation, Italy (Mutti et al.,

1978), demonstrating that these bed types commonly

occur in deep-water systems. Alternative explanations

for the generation of these dsandwichT beds are dis-

cussed by Haughton et al. (2003) and Talling et al.

(2004).

6. Conclusions

The behaviour of stratified gravity currents was

investigated using two-layer, laboratory flows com-

posed of aqueous glycerol solutions. In a set of

experiments the initial density and viscosity stratifi-

cation was systematically changed in a manner that

might occur in particle-laden currents with relatively

low to high sediment concentrations. It has been

shown previously that the vertical distribution of den-

sity and buoyancy profoundly affects the behaviour of

laboratory currents (Gladstone et al., 2004). Results

from this study show that the viscosity stratification

also has an important effect on flow behaviour. In

currents with relatively weak viscosity stratification

the high-concentration basal layer is driven by inertia

and propagates to the nose of the current, provided it

has a larger buoyancy than the upper layer. On the

other hand, in currents with relatively strong viscosity

stratification the high-concentration lower layer is

controlled by viscous forces and lags behind the

flow front regardless of its relative buoyancy. These

two flow types, with a relatively fast- and slow-mov-

ing lower layer, correspond to those with relatively

high and low Reynolds numbers, respectively. We

suggest that a transition in flow type occurs with the

onset of transitional and laminar flow conditions be-

cause of enhanced drag at the lower flow boundary. In

the present experiments this transition was observed at

concentrations of between 70% and 80% glycerol for

the lower layer.

The recorded temporal profiles of velocity and

concentration of currents with relatively weak and

strong viscosity stratification are different. Conse-

quently, stratified currents carrying particles are likely

to show different depositional histories and produce

deposits with varied characteristics. The experimental

results allow some speculation about the character of

stratified flow deposits. Currents with a relatively

fast-moving, high-concentration phase should deposit

beds with high-concentration flow deposits overlain

by those of more dilute flow. A current with a

relatively slow moving, high-concentration phase

should produce a bed with a tripartite structure with

a high-concentration flow deposit sandwiched be-

tween the deposits of more dilute flow. Deposits

with these bed structures are commonly observed in

ancient turbidite successions. Experiments on strati-


fied flows using particles is suggested as a fruitful

area of future work.

Acknowledgments

This research was funded by the United Kingdom

Natural Environmental Research Council and Con-

oco (now ConocoPhilips) through the Ocean Mar-

gins LINK scheme (Grant number NER/T/S/2000/

0106). Experiments were conducted in the Universi-

ty of Leeds, School of Earth Sciences, Fluid Dy-

namics Laboratory. Mark Franklin, Bob Bows, Gary

Keech, David Forgerty (School of Chemistry), Phil

Fields, Tony Windross are thanked for technical

support and members of Sedimentology Group for

assistance in running experiments. Jaco Baas,

Suzanne Leclair and an anonymous reviewer helped

to improve the original manuscript. Andy Hogg and

Charlotte Gladstone provided useful discussion.

Funding for the laboratory facilities used was pro-

vided by EPSRC (GR/R60843/01) and by a consor-

tium of oil companies including BG, BHP, Chevron

Texaco, Total, Exxon Mobil, ConocoPhillips, Ame-

rada Hess and Shell. We also acknowledge the

award of UK Natural Environment Research Council

grant GR3/10015 that funded development of the

UDVP system.

References

Allen, J.R, 1985. Principle of Physical Sedimentology. George

Allen & Unwin, London. 272 pp.

Altinakar, M.S., Graf, W.H., Hopfinger, E.J., 1996. Flow struc-

ture in turbidity currents. Journal of Hydraulic Research 34,

713–718.

Amy, L.A., 2000. Architectural Analysis of a Sand-rich Confined

Turbidite Basin; The Gres de Peıra Cava, South-East France.

PhD Thesis, University of Leeds, UK.

Amy, L.A., McCaffrey, W.D., Kneller, B.C., 2004. The influence of

a lateral basin-slope on the depositional patterns of natural and

experimental turbidity currents. In: Joseph, P., Lomas, S.A.

(Eds.), Deep-Water Sedimentation in the Alpine Basin of SE

France: New Perspectives on the Gres d’Annot and Related

Systems, Special Publication-Geological Society of London,

vol. 221, pp. 311–330.

Amy, L.A., Talling, P.J., Peakall, J., Wynn, R.B., Arzola Thynne,

R.G., 2005. Bed geometry used to test recognition criteria of

turbidites and (sandy) debrites. Sedimentary Geology 179,

161–172 (this issue) doi:10.1016/j.sedgeo.2005.04.007.

Argnani, A., Ricci Lucchi, F., 2001. Tertiary silicoclastic turbidite

systems of the Northern Apennines. In: Vai, G.B., Martini, I.P.

(Eds.), The Apennines and Adjacent Mediterranean Basins.

Kluwer Academic Press, pp. 327–350.

Arnott, R.W.C, Hand, B., 1989. Bedforms, primary structures and

grain fabric in the presence of suspended sediment rain. Journal

of Sedimentary Petrology 59, 1062–1069.

Barley, B., 1999. Deepwater problems around the world. Leading

Edge 18, 488–494.

Barnes, H.A., 1989. Shear-thickening (bdilatancyQ) in suspensions

of nonaggregating solid particles dispersed in Newtonian

liquids. Journal of Rheology 33, 329–366.

Best, J.L., Kirkbride, A.D., Peakall, J., 2001. Mean flow and

turbulence structure of sediment-laden gravity currents: new

insights using ultrasonic Doppler velocity profiling. In: McCaf-

frey, W.D., Kneller, B.C., Peakall, J. (Eds.), Particulate Gravity

Currents, Special Publication of the International Association of

Sedimentologists, vol. 31, pp. 159–172.

Bouma, A.H., Normark, W.R., Barnes, N.E., 1985. Submarine Fans

and Related Turbidite Systems. Springer Verlag, New York.

Branney, M.J., Kokelaar, P., 2002. Pyroclastic density currents and

the sedimentation of ignimbrites. Geological Society Memoir,

vol. 27. The Geological Society of London. 143 pp.

Buckee, C., Kneller, B., Peakall, J., 2001. Turbulence structure in

steady, solute-driven gravity currents. In: McCaffrey, W.D.,

Kneller, B.C., Peakall, J. (Eds.), Particulate Gravity Currents,

Special Publication of the International Association of Sedimen-

tologists, vol. 31, pp. 173–187.

Chadwick, A., Morfett, J., 1992. Hydraulics in Civil and Environ-

mental Engineering, 2nd edition Chapman and Hall.

Chen, Y.M., Pearlstein, A.J., 1987. Viscosity–temperature correla-

tion for glycerol–water solutions. Industrial & Engineering

Chemistry Research 26, 1670–1672.

Chikita, K., 1990. Sedimentation by river induced turbidity cur-

rents: field measurements and interpretation. Sedimentology 37,

891–905.

Choux, C.M.A., Baas, J.H., McCaffrey, W.D., Haughton, P.D.W.,

2005. Comparison of spatial–temporal evolution of experimental

particulate gravity flows at two different initial concentrations,

based on velocity, grain size and density data. Sedimentary Geol-

ogy 179, 49–69 (this issue) doi:10.1016/j.sedgeo.2005.04.010.

Coussot, P., 1997. Mudflow rheology and dynamics. Delft Hydrau-

lics, International Association for Hydraulic Research.

De Wit, P.J., 1992. Rheological measurements on artificial muds.

Report Number 9-92, Department of Civil Engineering., Delft

University of Technology, Delft, Netherlands.

Felix, M., 2002. Flow structure of turbidity currents. Sedimentology

49, 397–419.

Ferreira, J.M., Diz, H.M., 1999. Effect of solids loading on slip-

casting performance of silicon carbide slurries. Journal of the

American Ceramic Society 82, 1993–2000.

Fisher, R.V., 1995. Decoupling of pyroclastic currents: hazards

assessments. Journal of Volcanology and Geothermal Research

66, 257–263.

Garcıa, M.H., 1994. Depositional turbidity currents ladenwith poorly

sorted sediment. Journal of Hydraulic Engineering 120,

1240–1263.

http:\\dx.doi.org\doi:10.1016\j.sedgeo.2005.04.007

http:\\dx.doi.org\doi:10.1016\j.sedgeo.2005.04.010


Ghibaudo, G., 1992. Subaqueous sediment gravity flow deposits:

practical criteria for their field description and classification.

Sedimentology 39, 423–454.

Gladstone, C., Ritchie, L.J., Sparks, R.S.J., Woods, A.W., 2004. An

experimental investigation of stratified inertial gravity currents.


Hallworth, M.A., Huppert, H.E., 1988. Abrupt transitions in high-

concentration, particle-driven gravity currents. Physics of Fluids

10, 1083–1087.

Hampton, M.A., 1972. The role of subaqueous debris flow in

generating turbidity currents. Journal of Sedimentary Petrology

42, 775–793.

Haughton, D.W., Barker, S., McCaffrey, W.D., 2003. dLinkedTdebrites in sand-rich turbidite systems — origin and signifi-

cance. Sedimentology 50, 459–482.

Hickson, T.A., 1999. A study of deep-water deposition: constraints

on the sedimentation mechanics of slurry flows and high-con-

centration turbidity currents, and the facies architecture of a

conglomeratic channel overbank system. PhD Thesis, Stanford

University, San Francisco.

Hilton, V.C., 1994. Architecture of Deep-Marine Confined Sand-

stone Bodies, Eocene–Oligocene Gres d’Annot Formation SE

France. PhD Thesis, University of Leicester, UK.

Kneller, B.C., 1995. Beyond the turbidite paradigm: physical

models for deposition of turbidites and their implications for

reservoir prediction. In: Hartley, A.J., Prosser, C. (Eds.),

Characterization of Deep Marine Clastic Systems. Special

Publication, vol. 94. Geological Society of London, London,

UK, pp. 31–49.

Kneller, B.C., Buckee, C., 2000. The structure and fluid mechanics

of turbidity currents; a review of some recent studies and their

geological implications. Sedimentology 47, 62–94.

Kneller, B.C., McCaffrey, W.D., 2003. The interpretation of vertical

sequences in turbidite beds: the influence of longitudinal flow

structure. Journal of Sedimentary Research 73, 706–713.

Kreiger, I.M., Dougherty, T.J., 1959. A mechanism for non-New-

tonian flow in suspension of rigid spheres. Transactions of the

Society of Rheology 3, 137–152.

Lide, D.L., Baysinger, G., Berger, L.I., Goldberg, R.N., Kehiaian,

H.V., Kuchitsu, K., Lin, C.C., Rosenblatt, G., Smith, A.L.,

(Eds.), CRC Handbook of Chemistry and Physics, 84th Edition.

Available online (http://www.hbcpnetbase.com/).

Liu, K.F., Mei, C.C., 1990. Approximate equations for the slow

spreading of a thin sheet of Bingham plastic fluid. Physics of

Fluids. A, Fluid Dynamics 2, 30–36.

Lowe, D.R., 1982. Sediment gravity flows: II. Depositional models

with special reference to the deposits of high-concentration tur-

bidity currents. Journal of Sedimentary Petrology 52, 279–297.

Major, J.J., Pierson, T.C., 1992. Debris flow rheology: experimental

analysis of fine-grained slurries. Water Resources Research 28,

841–857.

Marr, J.G., Harff, P.A., Shanmugam, G., Parker, G., 2001. Experi-

ments on subaqueous sandy gravity flows: the role of clay and

water content in the flow dynamics and depositional structures.

Geological Society of America Bulletin 113, 1377–1386.

McCaffrey, W.D., Kneller, B., 2001. Process controls on the deve-

lopment of stratigraphic trap potential on the margins of con-

fined turbidite systems and aids to reservoir evaluation.

American Association of Petroleum Geologists Bulletin 85,

971–988.

McCaffrey, W.D., Choux, C.M., Baas, J.H., Haughton, P.D.W.,

2003. Spatio-temporal evolution of velocity structure, con-

centration and grain size stratification within experimental

particulate gravity currents. Marine and Petroleum Geology

20, 851–860.

Middleton, G.V., 1966. Experiments on density and turbidity cur-

rents: II. Uniform flow of density currents. Canadian Journal of

Earth Sciences 3, 627–637.

Middleton, G.V., 1993. Sediment deposition from turbidity currents.

Annual Review of Earth and Planetary Sciences 21, 89–114.

Mohrig, D., Whipple, K.X., Midhat, H., Ellis, C., Parker, G., 1998.

Hydroplaning of subaqueous debris flows. Geological Society

of America Bulletin 110, 387–394.

Mulder, T., Alexander, T., 2001. The physical character of submarine

density flows and their deposits. Sedimentology 48, 269–301.

Mutti, E., Nilsen, T.H., Ricci Lucchi, F., 1978. Outer fan depo-

sitional lobes of the Laga Formation (upper Miocene and

lower Pliocene), east-central Italy. In: Stanley, D.J., Kelling,

G. (Eds.), Sedimentation in Submarine Canyons, Fans and

Trenches, pp. 210–223.

Normark, W.R., 1989. Observed parameters for turbidity–current

flow in channel, reserve fan, Lake superior. Journal of Sedi-

mentary Petrology 59, 423–431.

Peakall, J., McCaffrey, W.D., Kneller, B., 2000. A process model for

the evolution, morphology, and architecture of sinuous submarine

channels. Journal of Sedimentary Research 70, 434–448.

Peakall, J., Felix, M., McCaffrey, W.D., Kneller, B., 2001. Partic-

ulate gravity currents: perspectives. In: McCaffrey, W.D., Knel-

ler, B.C., Peakall, J. (Eds.), Particulate Gravity Currents, Special

Publication of the International Association of Sedimentologists,

vol. 31, pp. 1–8.

Piper, D.J.W., Cochonat, P., Morrison, M.L., 1999. The sequence of

events around the epicentre of the 1929 Grand Banks earth-

quake: initiation of debris flows and turbidity current inferred

from sidescan sonar. Sedimentology 46, 79–97.

Postma, G., Nemec, W., Kleinspehn, K.L., 1988. Large floating

clasts in turbidites: a mechanism for their emplacement. Sedi-

mentary Geology 58, 47–61.

Ricci Lucchi, F., Valmori, E., 1980. Basin-wide turbidites in a

Miocene, over-supplied deep-sea plain: a geometrical analysis.


Richardson, J.F., Zaki, W.N., 1954. Sedimentation and fluidisation:

Part I. Transactions of the Institute of Chemical Engineers 32,

35–53.

Simpson, J.E., Britter, R.E, 1979. The dynamics of the head of a

gravity current advancing over a horizontal surface. Journal of

Fluid Mechanics 94, 477–495.

Stacey, M.W., Bowen, A.J., 1988. The vertical structure of turbid-

ity current: theory and observation. Journal of Geophysical

Research 93, 3528–3542.

Takeda, Y., 1991. Development of an ultrasound velocity profile

monitor. Nuclear Engineering and Design 126, 277–284.

Talling, P.J., Amy, L.A., Wynn, R.B., Peakall, J., Robinson, M.,

2004. Beds comprising debrite sandwiched within co-genetic

http:\\www.hbcpnetbase.com\


turbidite: origin and widespread occurrence in distal depositional

environments. Sedimentology 51, 163–194.

van Kessel, T., Kranenburg, C., 1996. Gravity current of fluid

mud on sloping bed. Journal of Hydraulic Engineering 122,

711–717.

Weimer, P., Link, M.H., 1991. Global petroleum occurrences in

submarine fans and turbidite systems. In: Weimer, P., Link,

M.H. (Eds.), Seismic Facies and Sedimentary Process of Sub-

marine Fans and Turbidite Systems. Springer-Verlag, New York,

pp. 9–67.

Wynn, R.B., Weaver, P.P.E., Masson, D.G., Stowe, D.A.V., 2002.

Turbidite depositional architecture across three inter-connected

deep-water basins on the Northwest African Margin. Sedimen-

tology 49, 669–695.


Sedimentary Geology 1

Combined measurements of velocity and concentration in

experimental turbidity currents

M. Felix*, S. Sturton, J. Peakall

School of Earth Sciences, University of Leeds, Leeds LS2 9JT, United Kingdom

Abstract

Three different sets of experimental turbidity currents were run in which velocity and concentration were measured

simultaneously, for several different heights above the bed. One set with cohesive sediment had an initial volumetric

concentration of 16% kaolinite, and the other two sets with non-cohesive sediment had concentrations of 28% and 4% silica

flour. Velocity was measured at 104–122 Hz using an Ultrasonic Doppler Velocimetry Profiler and concentration was measured

at 10 Hz using an Ultrasonic High Concentration Meter. The similarity of changes in velocity and concentration at the same

measurement heights are described and it is shown that the similarity depends on flow concentration and position in the flow.

The measurements are analysed using cross-correlations and wavelet analysis. Velocity measurements are compared with

analytical solutions for flow around a semisphere and flow around a half body. Measurements and analyses indicate that

turbulence is diminished by stratification, decoupling of regions where turbulence is generated and by reduction of vertical flow

in the turbidity currents.


Keywords: Turbidity currents; Laboratory experiments; Velocity and concentration measurements; Cohesive sediment; Non-cohesive sediment

1. Introduction

The temporal and spatial flow structure of turbid-

ity currents is the result of the interaction between

sediment and water (Kuenen, 1951; Lowe, 1982;

Altinakar et al., 1996; Felix, 2002). Sediment is the

driving force of these flows and it is transported as a

result of different mechanisms, such as particle–parti-

cle interactions, matrix strength and turbulence (Lowe,


doi:10.1016/j.sedgeo.2005.04.008

* Corresponding author.

E-mail address: [email protected] (M. Felix).

1982; Normark and Piper, 1991). Turbulence is gener-

ated by shear in the flow and can be damped by high

concentration (Kundu, 1990).

Details of the flow structure can influence deposits

and run-out distances of turbidity currents (Kneller and

Branney, 1995; Kneller and McCaffrey, 2003). Such

details include the location of turbulence generation

and the extent of similarity of spatial and temporal

changes in velocity and concentration, which are the

topics of the work presented here. At present, under-

standing of these details is only partial due to a lack of

suitable observations, both in the laboratory and in

nature.

79 (2005) 31–47

M. Felix et al. / Sedimentary Geology 179 (2005) 31–4732

Numerical modelling routinely includes both veloc-

ity and concentration (e.g., Hinze, 1960; van Andel and

Komar, 1969; Parker et al., 1986; Stacey and Bowen,

1988; Eidsvik and Brørs, 1989; Felix, 2001) and results

provide a good insight into flow structure and run-out

length. However, the link between velocity and con-

centration is generally prescribed by assuming that all

sediment has the same velocity as the water except for

the vertical direction where settling is included. Be-

cause of this assumption, laboratory experiments and

natural-scale observations can be useful.

Measurements in nature are generally for flows

with low sediment concentration as these are more

common and less destructive than high-concentration

flows. Combined velocity and turbidity observations

have been made in a number of studies, such as

Tesaker (1975), Hebbert et al. (1979), Fan (1986),

Chikita (1989), Chikita et al. (1991), Mitsuzawa et

al. (1993) and Khripounoff et al. (2003). These obser-

vations are limited to single heights in a flow or

present only single vertical profiles. Extensive tempo-

ral observations were made by Samolyubov (1986)

and Samolyubov and Bystrova (1994), but again,

these flows were of low concentration.

Combined measurements of velocity and concen-

tration in laboratory turbidity currents are presented

by Bonnefile and Goddet (1959), Tesaker (1969,

1975), Parker et al. (1987), Garcia and Parker

(1993), Garcia (1993, 1994), Altinakar et al. (1996),

Lee and Yu (1997) and Yu et al. (2000). These

studies present vertical profiles but do not show

changes with time. Many laboratory experiments

were small scale and flows were barely able to

keep sediment in suspension so the results are valid

for depositional flow stages only.

To address some of these limitations, combined

velocity and concentration measurements are pre-

sented here to look at both the temporal and spatial

flow structure of turbidity currents in short-lived,

rapid, lock-exchange flows. Results for flows of dif-

ferent concentration and sediment type (cohesive and

non-cohesive) are presented.

2. Methodology

Three different sets of lock-exchange experiments

were run where velocity and volumetric concentration

were measured simultaneously at the same height in

each run, with changing heights between runs in each

set. One set of five runs with cohesive sediment

(kaolinite) had initial volumetric concentration of

16% and a second set of five runs with non-cohesive

sediment (silica flour) had initial volumetric concen-

tration of 28%. Measurement heights for these two

sets were 23, 41, 94, 150 and 211 mm above the bed.

The third set of runs had non-cohesive sediment with

initial volumetric concentration of 4% and was mea-

sured at 11 heights: 4, 14, 23, 32, 41, 94, 150, 211,

261, 311 and 361 mm above the bed. The grain size of

the silica flour was d10=2 Am (10th percentile),

d50=9 Am (median grain size) and d90=27 Am(90th percentile) and that of the kaolinite was d10=1

Am, d50=6 Am and d90=35 Am.

The experiments were run in a 4.5-m-long-by-0.2-

m-wide-by-0.5-m-high perspex channel which was

open at the top and the downstream end. This chan-

nel was inserted in a larger (6 m by 0.5 m by 1.5 m)

glass-walled flume (Fig. 1). The insert channel rested

on a false floor, tilted at an angle of 58, positioned0.5 m above the bottom of the flume to create a

sump and to minimise end-wall reflections. The

flume was filled with tap water before each experi-

ment. A lock box was filled with a total volume of

120 l of water and sediment mixture as input for

each experiment and stirred using a mechanical

mixer. Stirring was stopped just before the opening

of the lock gate.

Downstream velocities were measured using a 2-

MHz Ultrasonic Doppler Velocimetry Profiler

(UDVP) in each run, giving a measuring frequency

of 104–122 Hz. For a further description see Best et

al. (2001), who also used this technique in turbidity

currents. The velocity probe was positioned 3.42 m

downstream of the lock gate in the centre of the

insert channel at the measurement heights mentioned

above. The water was 0.8 m deep at the measuring

location.

Concentration was measured using an Ultrasonic

High Concentration Meter (UHCM), which has a

measuring frequency of 10 Hz. The UHCM measures

obscuration between a sender and a receiver 10 mm

apart, outputting a voltage between 0 and 10 V. This

output signal is calibrated by measuring several sed-

iment water mixtures of known concentration, using

the same type of sediment as used in the experiments.

0.5

m

0.2

m

5o

688

mm

471

mm

568 mm

UHCM

3.4 m

120 L (not to scale)

mixer

Lock gate

4.5 m

UDVP rack for all measuring heights

Fig. 1. View of insert channel and lock box. The insert channel is open at the top and downstream ends and is placed in a larger flume filled with

tap water. The insert channel is placed 0.5 m above the base of the flume, creating a sump underneath. Note that velocity was recorded using a

single UDVP probe at a different height for each run. The lock box was not filled to capacity to prevent spilling during mixing.

M. Felix et al. / Sedimentary Geology 179 (2005) 31–47 33

The UHCM was positioned next to the UDVP probe

at the same height.

The results are analysed in several ways. Cross-

correlation of the velocity and concentration signals at

the same heights in each flow compares the shapes of

the normalised signals. Values of the cross-correlation

coefficient range from �1 (signals have the same

shape but have a phase shift of 1808) to +1 (signals

have the same shape and are in phase), with a value of

zero indicating no correlation at all. Low-amplitude

variations have little influence on the cross-correlation

coefficient which therefore measures how comparable

the waveforms and shape of the entire signals are

(e.g., Gubbins, 2004).

A more detailed comparison of the fluctuations is

provided by wavelet analysis, which shows the tempo-

ral variations of different scales for the concentration

and velocity signals (e.g., Farge, 1992; Kumar and

Foufoula-Georgiou, 1997). This type of analysis is

well suited for short-lived flows with a large range of

scales and allows comparison of the velocity and con-

centration signals with the same method. Large scales

represent large-scale motions while small scales repre-

sent rapid fluctuations caused by turbulence (Farge,

1992; Brunet and Collineau, 1997; Howell and

Mahrt, 1997). Paul, Haar and complex-valued Morlet

wavelets were used and showed the same features.

Because the Morlet wavelet results show the least

artefacts, only these are shown here. For long-duration

signals of constant frequency content, the relation be-

tween Morlet scale s and frequency f tends towards

s=2.5/f. The results are presented in scalograms where

scale is plotted against time and the amplitude of the

contribution of each scale to the signal is contoured.

Finally, velocities are compared with analytical

solutions for flow around different body shapes (semi-

sphere and half body, see, e.g., Hampton, 1972;

Kundu, 1990; McElwaine and Nishimura, 2001), to

show that velocity and concentration do not only

influence each other locally, but that the influence of

the near-bed flow extends high up into the flow.

Results are only shown for a measurement height of


211 mm, but other heights where concentration was

significantly lower than the concentration at a height

of 23 mm showed the same results.

Results from these analyses are combined with

previous theoretical work and observations to discuss

locations where turbulence is generated and the im-

portance for flow structure and deposit formation.

3. Experimental results

The results for the runs with 16% kaolinite are

shown in Fig. 2, for the 28% silica flour runs in

0 5 10 150

0.1

0.2

0 5 10 15

0 5 10 150

0.1

0.2

0 5 10 15

0 5 10 150

0.1

0.2

conc

entr

atio

n

0 5 10 15

0 5 10 150

0.1

0.2

0 5 10 15

0 5 10 150

0.1

0.2

time (sec0 5 10 15

Fig. 2. Concentration and downstream velocity profiles for flows of 16% k

of the plots, velocity profiles are thinner grey lines. All velocity profiles ar

bed while the bottom profile is closest to the bed; z is the height above

upstream. Time t =0 is the start of the measurement period when the lock

Fig. 3 and for the 4% silica flour runs in Figs. 4 and

5. Fig. 4 shows profiles at the same five heights as in

Figs. 2 and 3 while in Fig. 5 measurements at all 11

heights are used to construct a contour plot of con-

centration. The velocity profiles in Figs. 2–4 are

moving averages over 10 points, so they are of a

similar frequency as the concentration profiles and

allow easier comparison. Cross-correlation coeffi-

cients are shown in Table 1. All flows were fast

enough to keep sediment in suspension and no depo-

sition took place until the final part of the tail.

After the opening of the lock gate (t =0), all

velocity profiles show a high value followed by a

20 25 3020 25 30-500

0

500

1000z=211 mm

20 25 3020 25 30-500

0

500

1000z=150 mm

20 25 3020 25 30-500

0

500

1000u

(mm

/sec

)z=94 mm

20 25 3020 25 30-500

0

500

1000z=41 mm

20 25 30)

20 25 30-500

0

500

1000z=23 mm

aolinite. Concentration profiles are thick black lines near the bottom

e moving averages of 10 points. The top profile is furthest from the

the bed. Positive velocity is downstream, and negative velocity is

gate is lifted.

0 5 10 15 20 25 300

0.1

0.2

0.3

0 5 10 15 20 25 30

0

500

1000z=211 mm

0 5 10 15 20 25 300

0.1

0.2

0.3

0 5 10 15 20 25 30

0

500

1000z=150 mm

0 5 10 15 20 25 300

0.1

0.2

0.3

conc

entr

atio

n

0 5 10 15 20 25 30

0

500

1000

u (m

m/s

ec)

z=94 mm

0 5 10 15 20 25 300

0.1

0.2

0.3

0 5 10 15 20 25 30

0

500

1000z=41 mm

0 5 10 15 20 25 300

0.1

0.2

0.3

time (sec)0 5 10 15 20 25 30

0

500

1000z=23 mm

Fig. 3. Concentration and downstream velocity profiles for flows of 28% silica flour. Concentration profiles are thick black lines near the bottom

of the plots, and velocity profiles are the thinner grey lines. All velocity profiles are moving averages of 10 points. The top profile is furthest

from the bed while the bottom profile is closest to the bed; z is the height above the bed. Positive velocity is downstream, and negative velocity

is upstream. Time t =0 is the start of the measurement period and corresponds to the opening of the lock gate.


period of decrease. In some profiles, negative (up-

stream) velocity is present, especially at the higher

measuring positions. Velocity increases before the

concentration ramp-up, which indicates the water

that is displaced just before the arrival of the

turbidity current at the measurement location. Con-

centration always increases abruptly from zero. All

velocity and concentration profiles in the turbidity

currents show a number of different asymmetrical

peaks consisting of a rapid increase followed by a

more gradual decrease. The main flow duration

decreases from 10–20 s down to 2–5 s at the

highest measurement positions, and the number of

peaks decreases from about 4 to 1.

The cross-correlation coefficients are calculated for

the period of high concentration. Values are high

below 94 mm but somewhat lower for the two top-

most profiles, indicating a high degree of similarity

between the shapes of the two profiles without a shift

in phase. The cross-correlation coefficients indicate a

high degree of similarity, but in detail there are diffe-

rences (Figs. 2–4), with the greatest similarity for

(parts of) profiles with relatively low concentration,

for example, in the tails. Especially at a height of 94

mm, the profiles are very similar; the shapes, number

of peaks and duration of peaks are the same for the

velocity and concentration profiles in all three flows

(Figs. 2–4). Nearer to the bed, where concentration is

0 10 20 30 40 50 600

0.02

0.04

0 10 20 30 40 50 60

-200

0

200

400z=211 mm

0 10 20 30 40 50 600

0.02

0.04

0 10 20 30 40 50 60

-200

0

200

400z=150 mm

0 10 20 30 40 50 600

0.02

0.04

conc

entr

atio

n

0 10 20 30 40 50 60

-200

0

200

400

u (m

m/s

ec)

z=94 mm

0 10 20 30 40 50 600

0.02

0.04

0 10 20 30 40 50 60

-200

0

200

400z=41 mm

0 10 20 30 40 50 600

0.02

0.04

time (sec)0 10 20 30 40 50 60

-200

0

200

400z=23 mm

Fig. 4. Concentration and downstream velocity profiles for flows of 4% silica flour. Concentration profiles are thick black lines near the bottom

of the plots, and velocity profiles are thinner grey lines. All velocity profiles are moving averages of 10 points. The top profile is furthest from

the bed while the bottom profile is closest to the bed; z is the height above the bed. Positive velocity is downstream, and negative velocity is

upstream. Time t =0 is the start of the measurement period, when the lock gate is lifted.


highest, similarity is not as good (e.g., the 16% flow at

a measurement height of 23 mm; Fig. 2).

For all three flows at all measurement heights,

velocity decreases more gradually with time than

concentration (Figs. 2–4). For example, the temporal

decrease of the 16% flow concentration at 23 mm is

much more abrupt than that of velocity (Fig. 2). The

maximum velocity measured during the entire flow

period occurs in the second period of high velocity

values. The velocity maximum at any given time in

the 16% flow is measured at a height of 41 mm above

the bed. For the 28% and 4% flows, the maximum

velocity measured at a given time at the five heights is

either at 41 mm or at 23 mm above the bed. The

velocity maximum of all flows is therefore around

these heights. The velocity decreases gradually from

the bed towards the highest measurement positions

and only near the top does the maximum velocity

differ significantly from the maximum near-bed ve-

locity (Figs. 2–4). Near the bed, the concentration is

comparable to the input concentration but it decreases

towards the higher measurement positions.

In the 4% flows (Fig. 4), the concentration in the

first period of high values varies little at the different

measurement heights, but the vertical variation

increases away from the head. The highest concentra-

tion is in the second period of high values. The

maximum velocity is also behind the head and gets

closer to the bed with time. The interpretation of this

flow structure is that the amount of mixing is high in

0 5 10 15 20 25 30 350

0.05

0.1

0.15

0.2

0.25

0.3

0.35

time (sec)

z (m

)

0.04 0.04

0.03

0.02

0.01

0.005

0.010.020.03

0.04

0.01

0.01

0.0050.005

Fig. 5. Volumetric concentration contour plot using measurements at 11 different heights z above the bed in 11 different flows of 4% silica flour,

assuming that flow structure of each flow is comparable to the others. Contours are in volume fraction. Time t =0 corresponds to the opening of

the lock gate.


the head but decreases in the body, causing the sed-

iment to accumulate nearer the bed and provide a

larger driving force leading to higher velocity. In the

28% flows (Fig. 3), the concentration at the lowest

measurement height varies less than for the 4% flows

and concentration in the head decreases more towards

the top, causing these flows to be more vertically

stratified than the 4% flows. In the 16% flows (Fig.

2), the lowest profiles show an abrupt termination of

high concentration with time, but higher up, the pro-

files are more comparable to the 28% and 4% flows.

The vertical concentration decrease is largest in the

16% flows, the most stratified of the three.

In Fig. 5, all 11 measurement heights of the 4%

silica flour flows are included in contour plots of

concentration that show the temporal and vertical

spatial development. The head is about twice as

thick as the body and flow thickness is about 0.36

Table 1

Cross-correlation coefficients for the velocity and concentration

profiles in the turbidity currents at different measurement heights

23 mm 41 mm 94 mm 150 mm 211 mm

4% 0.9184 0.9343 0.8932 0.8117 0.8452

28% 0.8830 0.8102 0.8955 0.7852 0.6480

16% 0.9075 0.8503 0.8903 0.7670 0.6655

The profiles were normalised by the same maximum velocity and

concentration in each set of runs.

m (flow just reaches the highest measurement posi-

tion). The same set of peaks and troughs as seen in

Fig. 4 can also be seen in Fig. 5. The variations

extend downwards to the lowest measurement posi-

tion but the height towards which they extend up-

wards decreases with time. In the head, concentration

starts decreasing only above 0.15 m, but in later

periods of high concentration, the flow is more

stratified vertically.

4. Wavelet analysis

Results of the wavelet analysis are shown in Figs.

6–8 for the 16%, 28% and 4% flows, respectively. In

both velocity and concentration plots for all three

flows, the large scales normally have the highest

amplitude and these are also most persistent tempo-

rally. Small scales tend to vary more temporally and

have lower amplitude than large scales but this varies

depending on height in the flow and on concentration.

The small scales with the highest amplitudes are

present at the 94 mm height. There, both large and

small scales are of comparable amplitude and are

similar for velocity and concentration. Towards the

bed, small scales for concentration have lower ampli-

tude for flows of high concentration (Figs. 6 and 7)

but are still of relatively high amplitude for velocity.

Fig. 6. Scalograms for the 16% kaolinite runs using a Morlet wavelet. Left side figures are for velocity, and right side figures are for

concentration. The measurement heights are, from top to bottom, z =211, 150, 94, 41 and 23 mm, the same as shown in Figs. 2–4. Dark colours

indicate scales with high amplitude, and light colours indicate low amplitudes.


The scale with the highest amplitude in the velocity

profile is largest in the displaced water just in front of

the turbidity current and decreases towards the time of

onset of high concentration. This onset of high con-

centration is also associated with small-scale velocity

fluctuations. Near the top of the three flows, where

concentration is relatively low, the decrease with time

of the scale with the largest amplitude is smaller than

for lower measurement heights and not always notice-

able. In the concentration plots, the onset of high

concentration is seen by the pattern characteristic of

Morlet wavelet analysis of a box-shaped signal. The

abrupt decrease in near-bed concentration of the 16%

flow is similarly shown (Fig. 6).

5. Influence of near-bed flow on ambient water and

upper part of the flow

The two previous methods of analysis (cross-cor-

relation and wavelets) focussed on a comparison of

signals at the same height, but the mutual influence

of velocity and concentration extends throughout the

entire flow. The results for the 4% flows (Fig. 4)

Fig. 7. Scalograms for the 28% silica flour runs using a Morlet wavelet. Left side figures are for velocity, and right side figures are for




show several short periods of varying concentration,

while the results for the lowermost measurement

heights of the 16% flows (Fig. 2) show much

more constant concentration with only one dip pres-

ent a few seconds after the onset of high concentra-

tion. These different patterns of concentration lead to

different flow patterns in the upper parts of the

currents, as can be shown by comparing the present

velocity results to analytical solutions for flow

around two different body shapes (Fig. 9). The first

body shape used in the analysis is a semisphere,

which is half a sphere with its flat side on the bed

and a body of limited spatial extent. The second is a

half body which is a blunt body whose thickness is

zero at the bed and increases with distance from the

zero point. A half body is of much larger spatial

extent than the semisphere (e.g., Hampton, 1972;

Kundu, 1990; McElwaine and Nishimura, 2001).

The 4% flow shows several short periods of high

concentration with periods of low concentration in

between (Figs. 4 and 5), while the 16% flow shows a

long period of more consistently high concentration

Fig. 8. Scalograms for the 4% silica flour runs using a Morlet wavelet. Left side figures are for velocity, and right side figures are for




near the bed (Fig. 2), so the flow pattern in the 4%

turbidity current is expected to be more like that

around a semisphere while for the 16% flow the

pattern is expected to be more like that for a half

body.

The analytical solutions are strictly valid only

for ambient flow. However, the persistence of

wavelet scales between the displaced water in

front of the turbidity current and in the current

itself indicates a smooth transition between the

two. Additionally, the wavelet pattern at the highest

measurement position for all flows shows that the

influence of concentration on velocity is small at

these heights. For these reasons, the present exper-

imental results can be compared with the analytical

solutions.

-4 -3 -2 -1 0 1 2 30

1

2

3

4

z

flow lines around a half-body

0 1 2 3-1

-0.5

0

1

u

horizontal velocity above a half-body

-3 2 -1 0 1 2 30

0.5

1

1.5

2

z

flow lines around a semisphere

-3 -2 -1 0 1 2 30

0.5

1

x

u

horizontal velocity above a semisphere

-4 -3 -2 -1

0.5

Fig. 9. Analytical solutions for flow around a half body and flow

around a semisphere. Horizontal velocities at one height (see text

for equations) above the bodies are shown below the 2D vertical

cross sections through the bodies. Non-dimensional axis units.


For two dimensional flow around a half body, the

downstream velocity u is given (in polar coordinates)

by

u ¼ U þ mcos hð Þ=r;

where U [m/s] is the velocity of the ambient, m [m2/s]

is the material flux, r [m] is the radial coordinate and

h is the angular coordinate. For two-dimensional flow

around a semisphere, downstream velocity is given by

u ¼ U þ a2U=r2;

where a [m] is the radius of the hemisphere (see Fig.

9). These two analytical solutions are compared to the

velocity measurements in a reference frame moving

with the turbidity currents (Fig. 10). The 16% kaolin-

ite flow fits the pattern of flow around a half body,

consisting of initial velocity decrease, rapid increase

at the onset of the high concentration and final de-

crease. The first second of observed data does not fit

the trend, which is interpreted to be the result of the

initial disturbance due to lock gate opening and the

slumping phase of the flow (Simpson, 1997). The 4%

silica flour flow fits the pattern of symmetrical veloc-

ity increase and decrease of flow around a semisphere.

The 28% silica flour flow does not fit either pattern,

but is transitional between the two, with an initial

velocity decrease characteristic of the half body but

symmetrical flow as for the semisphere.

6. Discussion: implications for turbulence

generation and sedimentation

Turbulence is generated in turbidity currents by

shear throughout the entire flow, but three main loca-

tions have been recognised in previous studies. As in

all wall-bounded flows, turbulence is generated near

the bed, leading to high values close to the bed but

diminishing away from it. Turbulence is also generat-

ed near the top of the flow due to shear with the

ambient fluid, although this is less important for

flow in deep ambient than for laboratory experiments.

A third-generation region has been recognised in a

middle region of the flow, at a height just above the

height of the velocity maximum (Dallimore et al.,

2001; Buckee, 2000; Felix, 2002) but the cause of

turbulence at this location has not been well explained

yet. An explanation will be given below using obser-

vations from previous experimental and theoretical

work. The relative importance of turbulence generated

at the different heights varies. For example, a rough

bed will lead to high turbulence near the bed, while a

smooth bed will lead to less turbulence near the bed,

so that turbulence generated just above the height of

the velocity maximum may be more important.

Turbulence is highest near to where it is generated

and diminishes away from it. If turbulence generated

in different regions interacts, sediment can be distrib-

uted vertically throughout the flow more easily. A

lack of interaction has been described in the turbu-

lence minimum just below the height of the velocity

0 1 2 3 4 5 6 7 8 9 10-500

0

500

u (m

m/s

ec) 16 % kaol

0 1 2 3 4 5 6 7 8 9 10-500

0

500

1000

u (m

m/s

ec) 28 % sf

0 2 4 6 8 10 12 14 16 18 20-500

0

500

u (m

m/s

ec) 4 % sf

time (sec)

Fig. 10. Comparison of analytical solutions for flow around bodies with the measured velocity values for the three different flows. All

measurement heights are at 211 mm above the bed. Grey lines are the velocity measurements, continuous black lines are the analytical solutions

for flow around a half body, dashed black lines are the analytical solutions for flow around a semisphere, and the vertical dotted lines indicate

time of onset of high concentration.


maximum (e.g., Peakall et al., 2000), which is where

turbulence generated near the bed and turbulence

generated in the middle of the flow overlap at the

outer reaches of their influence.

In all three flows in the present experiments, the

wavelet analysis shows that the middle region (see

results for the 94 mm height), above the velocity

maximum, has high-amplitude velocity fluctuations

which will lead to high shear and therefore high

turbulence. For all three flows, the middle region, at

the transition between underflow and dragged upper

flow part, consistently generates turbulence, despite

the differences in concentration and stratification.

Stratification dampens turbulence because it is more

difficult for the denser fluid to be transported upwards

and for lighter fluid to move downwards. As a result,

generation of turbulence in the middle of the flow may

be expected to decrease for increasing stratification, in

contrast to the results shown. To understand sediment

distribution in the flows, it is important to understand

why waviness is generated in the middle of the flow.

The height above the velocity maximum with high

turbulence is the interface between the underflow and

the upper part of the flow which is dragged along.

Lock (1951) derived an analytical solution for the

velocity profile between an underflow and dragged

ambient and found a kink at this height. A similar

kink was found by Ippen and Harleman (1952) in

experiments with clay flows, at the upper surface of

the clay underflow. This kink increases shear and

generates additional turbulence, which was seen in

the numerical experiments of Felix (2002). Although

increased mixing due to turbulence may make this

transition smoother, the division remains clear even

after long flow duration as shown by the numerical

experiments of Felix (2004), where most of the flow

momentum remains concentrated in the underflow

below z1/2, where z1/2 is the height above the velocity

maximum where the velocity equals half the maxi-

mum velocity. Mixing will transport sediment up-

wards in the flow so that this transition height is not

obvious by visual inspection of flows only.

The interface above the height of the velocity max-

imum is not only the location for increased shear, but

also where internal waves will develop. If the transition

height between the lower and upper flow parts is also a

density interface, Kelvin Helmholtz waves may devel-

op (Kundu, 1990) and breaking of these waves leads to

turbulence. However, even if there is no or a very weak

density interface, waves will still develop due to shear


instability between two initially different fluid bodies

of the same density (Scorer, 1978) but distinguished by

a non-smooth change in the velocity profile at the

transition between the two flow parts. Samolyubov

and Bystrova (1994) applied the analytical solution of

Scorer describing this mechanism to their observations

of an underflow in a water storage reservoir. This

solution for flow of uniform density describes the

displacement pattern caused by wave formation. The

maximum displacement occurs at the transition be-

tween underflow and upper flow part, and, as for a

density interface, wave breaking may lead to turbu-

lence generation. For decaying waves, such as can be

expected for flows which are more vigorous at the front

than at the rear, this analytical solution shows displace-

ment as in Fig. 11, which is similar to that observed in

the 4% flows (Fig. 5). So for both high-concentration

flows with a density interface and for flows without a

density interface, waves will form. In most flows, both

mechanisms will occur simultaneously.

The results presented for a measurement height of

94 mm show very similar velocity and concentration

profiles (Figs. 2, 3 and 4) and comparable velocity

and concentration wavelet scales with both large and

small scales of high amplitude (Figs. 6, 7 and 8).

-5 0 50

0.5

1

1.5

2

2.5

3

3.5

dista

heig

ht

Fig. 11. Non-dimensional schematic plot of streamlines for the wavy moti

decaying waves. Solid line is the interface line. Method from Scorer (197

These results can be explained with the theoretical

models described above. Although turbulence is al-

ways generated at the middle height, irrespective of

stratification, concentration does have an influence on

the turbulence. For high-concentration flows, turbu-

lence does not have much influence on concentration

for the lowermost measuring positions (Figs. 2 and 3:

absence of small fluctuations, Figs. 6 and 7: absence

of small scales), where concentration is more constant

than for the low-concentration flows (Figs. 4 and 5).

Neither turbulence generated near the bed nor turbu-

lence generated near the middle influences the lower-

most measurement positions (which are, at 23 mm,

between these two heights where turbulence is gener-

ated), so their influence does not extend this far and

the two generation regions can be considered to be

decoupled, diminishing the vertical exchange of sed-

iment within the flow. Such decoupling, of course,

depends on the actual velocity, with high velocity

leading to high turbulence and more overlap between

different regions where turbulence is generated. Low

velocity leads to less turbulence and a decrease in

overlap of the regions. This means that decoupling

may result in a change from the front to the back of

the flow.

10 15 20nce

on formed at the interface of two different displacement regimes for

8).


The near-bed concentration varies less in higher

concentration flows and the flow pattern caused by

the concentration patterns in the lower part of the flow

is also different, as shown by the comparison of flow

around the different bodies (Fig. 10). For low-concen-

tration flows, the rapid alternation between periods of

(relatively) high and low concentration (see, e.g., Figs.

4 and 5) leads to many periods of upwards and down-

wards flow, as indicated by flow in front of and behind

the semisphere (Fig. 10). For high-concentration flow,

vertical motion is more restricted and flow is more like

flow around and parallel to a half body (Fig. 10). The

imposed forcing up and down of the water for low-

concentration flows leads to high shear and generation

of turbulence. For high-concentration flows, the im-

posed up and down flow decreases as the flow is more

parallel to the long lower part of the flow, leading to a

decrease in shear caused by vertical motion and result-

ing in a decrease of turbulence. It will therefore be

more difficult to mix and suspend sediment in the

higher regions of the flow.

The results of wavelet analysis and flow around a

body therefore show that in higher concentration

flows, turbulence is not only damped by stratification,

but also by decoupling of turbulence generation

regions which inhibits vertical mixing of sediment,

and by a change in the flow patterns in the dragged

upper part of the flow, again inhibiting mixing. These

flow patterns are shown schematically in Fig. 12. The

results now allow a short discussion of the implication

of these flow patterns for deposition, but this discus-

sion necessarily remains speculative as the flows in

the present experiments were non-depositional, so the

implications cannot be confirmed from them. Results

from numerical modelling or observations in deposi-

tional natural-scale flows would be most suitable for

this, being of the right spatial and temporal scales.

The experimental results show the similarity be-

tween the velocity and concentration profiles, with

variations depending on height in the flow and on

flow concentration. When profiles are similar, a

change in either velocity or concentration will be

followed rapidly by a change in the other. Deposition

takes place from sediment transported near the bed

and rapid or slow changes in concentration will result

in rapid or slow changes in deposits and bedforms.

Prediction of deposit changes based on velocity

changes (e.g., Kneller and Branney, 1995; Kneller

and McCaffrey, 2003) depends on the reaction of

concentration to velocity changes.

For low-concentration flows, different turbulence

generation regions interact (see Fig. 12) and sediment

will be transported vertically throughout the flow

relatively easily. The vertical transport of sediment

near the bed is influenced both by turbulence gener-

ated near the bed and by turbulence generated just

above the height of the velocity maximum. A change

in velocity will only be slowly followed by a change

in concentration as near-bed concentration will be

affected relatively little by a change in turbulence

generated near the bed which will be compensated

by turbulence generated just above the velocity max-

imum. Bed thickness will change gradually and

changes in bedforms will also be gradual in the down-

stream direction.

For increasing concentration, the near-bed region

will become decoupled from the middle region (see

Fig. 12) and sediment is transported less easily

vertically. Near-bed concentration is mostly influ-

enced by turbulence generated at the bed with little

influence from turbulence generated above the ve-

locity maximum. A change in velocity and bed

shear will accordingly have a relatively larger

change in concentration than for low-concentration

flows. The resulting deposits will therefore record

smaller changes in velocity more accurately than

low-concentration flows. This is the case for both

deposition and erosion. Deposit thickness can vary

rapidly if the underlying topography causes the

flow velocity to increase or decrease suddenly and

bedforms can also change rapidly from one location

to another.

The two wave generation mechanisms described

above show how turbidity currents may be wavy and

the resulting velocity variations lead to local waxing

and waning behaviour. Although in the present lab-

oratory flows the velocity varies on a scale of sec-

onds, such behaviour has been observed in natural-

scale flows (Samolyubov and Bystrova, 1994) to be

on the order of minutes to hours. This might be

recorded in the deposit through thick laminae.

Lowe (1982) described such lamination (traction

carpets) in deposits of high-concentration flows

being formed as the result of rapid freezing of the

near-bed sediment, but this will only happen for

near-bed concentration close to packing concentra-

High density with stiff layer near the bed that can deform: pluglike flow.

Intermediate density with high density layer near the bed.

TKE

ULow density overall.

LD

dense, mixed

LD

LD

coherent

Fig. 12. Schematic diagram for the three different flows. Left hand column shows a 2D side view of the currents with the different concentration

regimes and the ambient flow around the current (thick black arrow). Right hand column show regions of turbulence generation and their

interaction as well as vertical turbulent kinetic energy (TKE) profiles, linked to height in vertical velocity profiles. LD=low-density part of the

flow. The figure is based on present results and previous theoretical and observational work. See text for discussion.


tion. Lower concentration flows can lead to similar

laminations as a result of temporal variations in flow

structure. In such waxing and waning flows, vertical

grain size trends such as inverse grading are more

likely to be preserved in deposits from high-concen-

tration flows if it is assumed that the increase or

decrease in velocity is accompanied by equivalent

changes in transported grain size. In high-concentra-

tion flows, deposition can be so rapid that smaller

grain sizes already deposited are not eroded in a

waxing flow because relatively little turbulence is

available to transport it away from the bed.

7. Conclusions

Combined measurements of velocity and concen-

tration in experimental turbidity currents show that the

similarity of temporal and vertical spatial changes in

velocity and concentration in turbidity currents

depends on bulk concentration of the flow and on

position in the flow. Similarity is high near the middle

of the flow at the transition height between underflow

and ambient. For high concentration, the similarity

decreases, especially near the bed. Despite these dif-

ferences, cross-correlation coefficients of the two sig-


nals are high, indicating the similarity of the overall

signals and the absence of a phase shift.

Wavelet analysis shows that all profiles have one

large scale of high amplitude which is comparable for

velocity and concentration, while smaller scales gener-

ally have lower amplitude. Small and large scales at the

transition height between underflow and upper part of

the flow are of comparable amplitude as a result of

turbulence generation at this height. For high-concen-

tration flows, turbulence generation regions near the

bed and at the transition height become decoupled

which inhibits vertical mixing of sediment.

Temporal changes in the low-concentration flow

are rapid with several periods of alternating low and

high concentration. This pattern results in flow com-

parable to flow around a semisphere. For high-con-

centration flows, the concentration near the bed varies

more slowly, resulting in a flow pattern comparable to

flow around a half body. Mixing of sediment is re-

duced for high-concentration flows by the decreased

amount of imposed vertical motion as the near-bed

sediment concentration is more constant than for low-

concentration flows.

Acknowledgements

This research was funded by the UK Engineering

and Physical Sciences Research Council, Grant GR/

R60843/01. The UDVPs were funded by UK Natural

Environment Research Council Grant GR3/10015.

Mark Franklin and Gareth Keevil are thanked for

help in the laboratory. Reviews by Yu’suke Kubo,

Jeff Parsons and Jaco Baas helped to significantly

improve the clarity of the paper.

References

Altinakar, M.S., Graf, W.H., Hopfinger, E.J., 1996. Flow structure

in turbidity currents. Journal of Hydraulic Research 34 (5),

713–718.

Best, J., Kirkbride, A., Peakall, J., 2001. Mean flow and turbulence

structure of sediment-laden gravity currents: new insights using

ultrasonic Doppler velocity profiling. In: McCaffrey, W.D.,


IAS Special Publications, vol. 31, pp. 159–172.

Bonnefile, R., Goddet, J., 1959. Etude des courants de densite en

canal. Proceedings Eight Congress of the International Associ-

ation for Hydraulic Research, vol. 2, pp. 14-C-1–14-C-29.

Subjects C and D.

Brunet, Y., Collineau, S., 1997. Wavelet analysis of diurnal and

nocturnal turbulence above a maize crop. In: Kumar, P., Fou-

foula-Georgiou, E. (Eds.), Wavelet Analysis for Geophysical

Applications, pp. 129–150.

Buckee, C.M., 2000. Mean flow and turbulence structure in exper-

imental gravity currents. Ph.D. Thesis, University of Leeds,

United Kingdom.

Chikita, K., 1989. A field study on turbidity currents initiated from

spring runoffs. Water Resources Research 25 (2), 257–271.

Chikita, K., Yonemitsu, N., Yoshida, M., 1991. Dynamic sedimen-

tation processes in a glacier-fed lake, Peyto Lake, Alberta,

Canada. Japan Journal of Limnology 52 (1), 27–43.

Dallimore, C.J., Imberger, J., Ishikawa, T., 2001. Entrainment and

turbulence in saline underflow in Lake Ogawara. Journal of

Hydraulic Engineering 127 (11), 937–948.

Eidsvik, K.J., Brørs, B., 1989. Self-accelerated turbidity current

prediction based upon (k–e) turbulence. Continental Shelf Re-

search 9 (7), 617–627.

Fan, J., 1986. Turbid density currents in reservoirs. Water Interna-

tional 11, 107–116.

Farge, M., 1992. Wavelet transforms and their applications to

turbulence. Annual Review of Fluid Mechanics 24, 395–457.

Felix, M., 2001. A two-dimensional numerical model for a turbidity

current. In: McCaffrey, W.D., Kneller, B.C., Peakall, J. (Eds.),

Particulate Gravity Currents, IAS Special Publications, vol. 31,

pp. 71–81.


49, 397–419.

Felix, M., 2004. The significance of single-value variables in tur-

bidity currents. Journal of Hydraulic Research 42 (3), 323–330.

Garcıa, M.H., 1993. Hydraulic jumps in sediment-driven bot-

tom currents. Journal of Hydraulic Engineering 119 (10),

1094–1117.

Garcia, M.H., 1994. Depositional turbidity currents laden with

poorly sorted sediment. Journal of Hydraulic Engineering 120

(11), 1240–1263.

Garcia, M., Parker, G., 1993. Experiments on the entrainment of

sediment into suspension by a dense bottom current. Journal of

Geophysical Research 98 (C3), 4793–4807.

Gubbins, D., 2004. Time Series Analysis and Inverse Theory for

Geophysicists. Cambridge University Press, Cambridge.

Hampton, M.A., 1972. The role of subaqueous debris flow in

generating turbidity currents. Journal of Sedimentary Petrology

42 (4), 775–793.

Hebbert, B., Imberger, J., Loh, I., Patterson, J., 1979. Collie river

underflow into the Wellington reservoir. ASCE Journal of the

Hydraulics Division 105 (5), 533–545.

Hinze, J.O., 1960. On the hydrodynamics of turbidity currents.

Geologie en Mijnbouw, 39e jaargang, 18–25.

Howell, J.F., Mahrt, L., 1997. An adaptive decomposition: applica-

tion to turbulence. In: Kumar, P., Foufoula-Georgiou, E. (Eds.),

Wavelet Analysis for Geophysical Applications, pp. 107–128.

Ippen, A.T., Harleman, D.R.F., 1952. Steady-state characteristics of

subsurface flow. Gravity Waves. National Bureau of Standards

Circular 521, pp. 79–93.


Khripounoff, A., Vangriesheim, A., Babonneau, N., Crassous, P.,

Dennielou, B., Savoye, B., 2003. Direct observation of intense

turbidity current activity in the Zaire submarine valley at 4000 m

water depth. Marine Geology 194, 151–158.

Kneller, B.C., Branney, M.J., 1995. Sustained high-density turbidity

currents and the deposition of thick massive sands. Sedimentol-

ogy 42 (4), 607–616.

Kneller, B.C., McCaffrey, W.D., 2003. The interpretation of vertical

sequences in turbidite beds: the influence of longitudinal flow

structure. Journal of Sedimentary Research 73, 706–713.

Kuenen, Ph.H., 1951. Properties of turbidity currents of high den-

sity. Turbidity Currents and the Transportation of Coarse Sedi-

ments to Deep Water—a Symposium, Society of Economic

Paleontologists and Mineralogists Special Publication, vol. 2,

pp. 14–33.

Kumar, P., Foufoula-Georgiou, E., 1997. Wavelet analysis for geo-

physical applications. Reviews of Geophysics 35 (4), 385–412.

Kundu, P.K., 1990. Fluid Mechanics. Academic Press Ltd, London.

Lee, H.-Y., Yu,W.-S., 1997. Experimental study of reservoir turbidity

current. Journal of Hydraulic Engineering 123 (6), 520–528.

Lock, R.C., 1951. The velocity distribution in the laminar boundary

layer between parallel streams. Quarterly Journal of Mechanics

and Applied Mathematics IV (1), 42–57.

Lowe, D.R., 1982. Sediment gravity flows II. Depositional models

with special preference to the deposits of high-density turbidity

currents. Journal of Sedimentary Petrology 52 (1), 279–297.

McElwaine, J., Nishimura, K., 2001. Ping-pong ball avalanche

experiments. In: McCaffrey, W.D., Kneller, B.C., Peakall, J.

(Eds.), Particulate Gravity Currents, IAS Special Publications,

vol. 31, pp. 135–148.

Mitsuzawa, K., Momma, H., Fukasawa, M., Hotta, H., 1993. Ob-

servation of deep sea current and change of bottom shapes in the

Suruga trough. Proceedings of the Conference on Oceans ’93,

Part 3, pp. 149–154.

Normark, W.R., Piper, D.J.W., 1991. Initiation processes and flow

evolution of turbidity currents: implications for the depositional

record. In: Osborne, R.H. (Ed.), From Shoreline to Abyss,

Society of Economic Paleontologists and Mineralogists Special

Publication, vol. 46, pp. 207–230.

Parker, G., Fukushima, Y., Pantin, H.M., 1986. Self-accelerating

turbidity currents. Journal of Fluid Mechanics 171, 145–181.

Parker, G., Garcia, M., Fukushima, Y., Yu, W., 1987. Experiments

on turbidity currents over an erodible bed. Journal of Hydraulic

Research 25 (1), 123–147.

Peakall, J., McCaffrey, W., Kneller, B., 2000. A process model

for the evolution, morphology, and architecture of sinuous

submarine channels. Journal of Sedimentary Research 70

(3), 434–448.

Samolyubov, B.I., 1986. Interaction between shear layers and the

formation of inversion structures in benthic stratified flow.

Oceanology 26 (6), 695–703.

Samolyubov, B.I., Bystrova, N.A., 1994. The structure of the den-

sity current and of internal waves in its depth. Moscow Univer-

sity Physics Bulletin 49 (1), 77–83.

Scorer, R.S., 1978. Environmental aerodynamics, Elliswood Hor-

wood limited publishers (Chichester), section 5.10, pp. 173–181.

Simpson, J.E., 1997. Gravity Currents in the Environment and the

Laboratory. Cambridge University Press, Cambridge.

Stacey, M.W., Bowen, A.J., 1988. The vertical structure of density

and turbidity currents: theory and observations. Journal of Geo-

physical Research 93 (C4), 3528–3542.

Tesaker, E., 1969. Uniform turbidity current experiments. Thirteenth

Congress of the International Association for Hydraulic Re-

search, vol. 2, pp. 1–8. Subject B.

Tesaker, E., 1975. Modelling of suspension currents. Symposium on

Modeling Techniques, vol. II, pp. 1385–1401.

van Andel, T.H., Komar, P.D., 1969. Ponded sediments of the Mid-

Atlantic ridge between 22 degrees and 23 degrees north latitude.

Geological Society of America Bulletin 80 (7), 1163–1190.

Yu, W.-S., Lee, H.-Y., Hsu, S.M., 2000. Experiments on deposition

behaviour of fine sediment in a reservoir. Journal of Hydraulic

Engineering 126 (12), 912–920.



Comparison of spatio–temporal evolution of experimental

particulate gravity flows at two different initial concentrations,

based on velocity, grain size and density data

C.M.A. Choux a,*, J.H. Baas a, W.D. McCaffrey a, P.D.W. Haughton b

aEarth Sciences, School of Earth and Environment, University of Leeds, Leeds, LS2 9JT, UKbDepartment of Geology, University College Dublin, Belfield, Dublin 4, Ireland

Received 1 May 2004; accepted 6 April 2005

Abstract

Flume experiments were conducted to investigate the spatio–temporal structure of subaqueous particulate gravity flows with

an initial concentration of 14% by volume. Time series of downstream flow velocity and its calculated degree of turbulence,

median grain size and sediment concentration at different positions along the path of nominally identical flows are analysed and

combined to constrain the spatio–temporal evolution of a single idealised flow. Comparison of the 14% flow with a flow of 5%

initial concentration reveals similarities in the basic spatio–temporal structure of velocity, turbulence, grain size and concen-

tration. Both flow types exhibit a velocity maximum at about 1 /3 of the flow height above the flume floor. At that level,

velocity decreases slowly in the flows’ body and more rapidly in their tails. Moreover, turbulence intensity is highest in the head

and at the base of the flows, whereas the level of maximum velocity and the tail of the flows typically are weakly turbulent. The

zones of high turbulence are associated with shear at the front and base of the gravity flows. The flow of 5% and 14% initial

concentration also agree in stratification patterns of median grain size and concentration. Grain populations are relatively well

mixed in the head, show normal grading in the main part of the body and normal to inverse grading in the rear of the body and

tail. The inverse grading is thought to originate from particles transported from the head upward and backward into the body of

the flows, where they subsequently settle. The main difference between the flow of 5% and 14% initial concentration is that the

higher-density flows appear to develop from a jet into a turbidity current closer to the inception point than the lower-density

flow. This difference is interpreted from dimensionless vertical profiles of the flow parameters: horizontal velocity, concen-

tration and grain size distribution. In the turbidity current phase of both flows, the dimensionless variables collapse well. This

indicates that the flows behave in a dynamically similar manner and inspires confidence that the dimensionless variables can be

used to predict the dynamic behaviour of particulate gravity flows across the measured concentration range in the flume, which

due to dilution/sedimentation effects, was from ~7 to b1 vol.% concentration.


Keywords: Turbidity current; Flume experiments; Horizontal velocity; Root-Mean-Square velocity; Concentration; Grain size

* Corresponding author.

0037-0738/$ - s

doi:10.1016/j.se

E-mail addre

79 (2005) 49–69

ee front matter D 2005 Elsevier B.V. All rights reserved.

dgeo.2005.04.010

ss: [email protected] (C.M.A. Choux).

C.M.A. Choux et al. / Sedimentary Geology 179 (2005) 49–6950

1. Introduction

Particulate gravity currents, both subaerial and

subaqueous, exhibit a wide range of density. Turbidity

currents are an intrinsic part of the spectrum of sub-

aqueous sediment gravity flows. Although they too

encompass a wide range of concentration, for several

decades now, turbidity currents have been consistent-

ly described as sediment-laden gravity-driven flows in

which the sediment is supported principally by fluid

turbulence (Sanders, 1965; Middleton and Hampton,

1973; Middleton, 1993; Simpson, 1997; Shanmugam,

1997). Nevertheless, grain support mechanisms other

than fluid turbulence may co-occur in turbidity cur-

rents, such as hindered settling, particle–particle inter-

actions and buoyancy enhancement (Hiscott, 1994;

Mulder and Alexander, 2001; also see reviews by

Stow et al., 1996 and Kneller and Buckee, 2000).

The contribution of these mechanisms to grain sup-

port is highly dependent on the local concentration of

suspended sediment in the flow, and thus the relative

importance of these mechanisms may change if flows

change their concentration structure as they develop.

In most cases, turbidity currents do indeed evolve in

concentration as they flow, either through sediment

erosion and entrainment (e.g., Pantin, 1979), or

through deposition and entrainment or detrainment

of ambient water (Simpson, 1997 and references

therein). This triggers the question of how flows of

differing initial concentration compare in terms of

internal grain size distribution, concentration and ve-

locity structure as they develop. Do initially dense,

depositional flows, propagating for sufficient time to

become dilute, show the same basic dynamical be-

haviour as initially dilute flows? Are high concentra-

tion turbulent flows viable as a long range transport

mechanism, or are high concentrations only devel-

oped transiently, during sediment entrainment and/or

deposition?

A large volume of experimental work has been

undertaken in order to analyse the role of particle

concentration in turbidity current behaviour and struc-

ture, as well as the geometry and internal structure of

their deposits (e.g., Kuenen, 1966; Middleton, 1967;

Britter and Simpson, 1978; Luthi, 1980; Laval et al.,

1988; Middleton and Neal, 1989; Altinakar et al.,

1990; Bonnecaze et al., 1993; Garcia and Parsons,

1996; Gladstone et al., 1998; Hallworth and Huppert,

1998; Kneller et al., 1999; Stix, 2001; Choux and

Druitt, 2002; McCaffrey et al., 2003; Baas et al.,

2004; Al-Ja’Aidi et al., 2004; see also reviews by

Edwards, 1993; Middleton, 1993; Kneller and

Buckee, 2000, and Shanmugam, 2000). Despite the

value of these experimental works, the results cannot

be used directly to answer questions regarding the

spatio–temporal evolution of natural particulate grav-

ity currents. This requires a detailed characterisation

of both the spatial and temporal evolution of the

properties of experimental particle-driven flows in

terms of velocity, granulometric and concentration

structure, across a range of concentrations. Until re-

cently, however, only temporal (time series) data were

collected (see discussion in Peakall et al., 2001), with

experiments therefore focussing on flow unsteadiness

rather than flow non-uniformity.

Best et al. (2001) used 4-MHz ultrasonic Doppler

velocity profiling (UDVP) to quantify, for the first

time, the spatial and temporal evolution of mean

flow and turbulence structure of sediment-laden par-

ticulate flows. However, concentration and grain size

data were not collected, and the length over which

flow evolution was characterised (up to 85 mm) was

relatively small compared to the scale of the flows.

By coupling the UDVP method with a siphoning

technique and sampling several identical depositing

flows at different locations, McCaffrey et al. (2003)

were the first to produce a detailed description of the

spatio–temporal evolution of a particulate current of

5% initial concentration in terms of instantaneous

velocity, grain size and concentration. Their study

was limited to flows of a single initial starting

concentration.

Computer modellers also endeavour to shed new

lights on describing turbidity current structure (e.g.,

Stacey and Bowen, 1988; Zeng and Lowe, 1997;

Mulder et al., 1998). With the help of a non-depth

averaged model, using a multiphase flow approach,

and particle–particle interaction, incorporating the tur-

bulence model of Mellor and Yamada (1982), Felix

(2001, 2002) produced a 2-D, vertical plane, numer-

ical model that simulates unsteady flow behaviour in

terms of velocity, turbulence, grain size and concen-

tration distribution. However, to date his results have

been tested only against a small number of incomplete

historical flow data, because of the lack of suitable

experimental data (Felix, 2002).

C.M.A. Choux et al. / Sedimentary Geology 179 (2005) 49–69 51

The experimental data of McCaffrey et al. (2003)

can only be used to validate numerical models of

flows in which particle–particle interactions can be

ignored, as the initial particle concentration (5% by

volume) was below the 9 vol.% threshold proposed by

Bagnold (1966) above which these effects become

significant. In the present paper, a new set of experi-

ments is presented, using the same experimental set-

ting as McCaffrey et al. (2003), but for a higher initial

concentration (14%), at which particle–particle inter-

actions should influence flow behaviour, via moderate

turbulence suppression (Middleton and Hampton,

1976; Lowe, 1982). Thus the aim of this work is to

enable the spatio–temporal evolution of the internal

structure of relatively high concentration turbidity

currents (14%, hereafter referred to as high-density

flow) to be compared with that of more dilute flows

(5%, hereafter referred to as low-density flow) in

terms of the vertical gradients in instantaneous hori-

zontal velocity, grain size and concentration distribu-

tion, as well as turbulence structure.

Accordingly, the experimental set up of McCaffrey

et al. (2003) are described, and the key results are

summarised. The flow structure of a turbidity current

with 14% initial sediment concentration is then de-

tailed and compared with the 5% flows results of

McCaffrey et al. (2003). Subsequently, original analy-

sis of the turbulence structure of flows of both 5% and

14% initial concentration (as expressed by root-mean-

square–RMS–velocities) is presented. Dimensionless

parameters are established with the aim of comparing

the low- and high-density flows in more detail and

expanding the results to a wider range of initial sed-

iment concentrations.

2. Previous related work

Based on experimental data, the approach of

McCaffrey et al. (2003) allowed for the first time,

the structure of a turbidity current to be reconstructed

at any position and time, for the parameters stream-

wise velocity, grain size and suspended sediment

concentration. In a series of flume experiments,

McCaffrey et al. (2003) generated subaqueous partic-

ulate gravity flows through release of a 30 l suspension

of non-cohesive material (silica flour) at an initial

concentration of 5% by volume. They measured si-

multaneously the temporal evolution of the vertical

stratification in streamwise velocity, flow concentra-

tion and grain size distribution as the entire flow

passed a measurement location. A series of five nom-

inally identical flows were run, with measurements

repeated at five different locations along the flow

path. The results were then combined to constrain

the spatio–temporal evolution of a single idealised

flow. The inbound jet transformed into a gravity-

driven current at a distance between 1.32 and 2.64

m from the reservoir, and thereafter developed under

its own internal action.

The experimental depositional particulate gravity

currents of McCaffrey et al. (2003) were non-uniform,

i.e., their structure varied spatially (see Allen, 1985

and Kneller and Branney, 1995), indicating that it

would be erroneous to interpret time series data of

such flows in terms of longitudinal flow structure, as

commonly done in the existing literature. For exam-

ple, the velocity data showed that the flow duration

increased downstream, as the flow stretched out. The

concentration data showed that in proximal locations,

the rate of decrease of concentration was high, indi-

cating rapid sedimentation, whereas in distal locations

the rate of decrease was more gradual.

The transition between the head and the body of

each nominally identical turbidity current was de-

scribed by a sharp decrease in the maximum velocity

and median grain size, whereas the transition between

the body and the tail was well defined by a decrease in

the concentration. The velocity maximum was located

at approximately one third of the flow’s height from

its base. An interesting normal to inverse vertical

pattern in grading observed in the grain size distribu-

tion of material suspended in the flow’s body was

linked to the presence of coarse sediment inferred to

have been swept upwards and backwards over the

head then falling passively into the upper part of the

flow.

3. Experimental set up

The experimental set up in this study was the same

as that used by McCaffrey et al. (2003), with the sole

difference that the concentration of the initial suspen-

sion was increased from 5% to 14% by volume (i.e.,

with an initial suspension density of 1231 kg m�3


compared to 1082 kg m�3). The flume in which the

five nominally identical flows were run was 10 m

long, 0.3 m wide and 0.3 m deep (Best et al., 2001),

with an overhead reservoir containing 30 l of suspen-

sion (Fig. 1a). A homogeneous suspension in water of

silica flour particles (density: 2650 kg m�3) was

created in the reservoir, and kept well mixed by a

mechanical stirrer. The particle size ranged from

b0.01 to c60 microns, with a median grain size

(D50) of about 8 Am (Fig. 1b). At the start of each

experiment, the stirrer was stopped and a sealing

stopper at the bottom of the reservoir was removed

swiftly. Time series of the low-density experiments of

McCaffrey et al. (2003) taken 0.04 m downstream

from the reservoir outlet (their Flows 1 and 2, illus-

trated in their Fig. 4) showed that the inbound flows

were steady in terms of velocity, grain size and con-

centration for 21.5 s before swiftly decelerating. From

this it may be inferred that the suspension in the

reservoir remained essentially uniform as it drained.

This is probably due to the inherited turbulence from

the mixer, plus any turbulence generated by shear

against the reservoir walls as the suspension flowed

out and into the flume. Although similar outlet time

series were not collected for the high-density experi-

ments reported here, it is inferred that the input to the

flume was essentially steady in this case too. The

suspension drained into the water-filled flume through

a circular pipe of 0.063 m diameter, emptying the

reservoir in about 21.5 s at constant discharge (cf.,

McCaffrey et al., 2003) and forming a particulate

gravity current that propagated along the length of

the flume. The flow was sampled by an array of

instruments all positioned at the same location (Figs.

1 and 2). Due to the intrusiveness of the data acqui-

sition method, each flow could be measured at one

location only. Thus the reservoir was shifted upstream

Fig. 1. a) Experimental set up and

by an interval distance of 1.32 m between each of the

five nominally identical flows, increasing the distance

between the entry point of the flow and the measure-

ment point (Fig. 1a). The height of the flow’s head

was fairly constant at about 0.08–0.085 m, when

reaching the sampling devices, at all locations. Up-

ward flow motion was observed in the head of the

flow as the flow propagated; this upward moving fluid

was then forced back to horizontal by the ambient

water swept over the front of the head (Fig. 2). At

Location 1, the Reynolds number, calculated using

average values within the head, was 2�104, and so

was well within the turbulent regime. Data were ac-

quired from the time of flow inception until upstream-

propagating, solitary waves (e.g., Pantin and Leeder,

1987; Edwards, 1993), generated by the reflection of

the inbound flow from a distal overflow weir (Fig. 1)

passed the array of instruments.

Two sets of instruments were used at each mea-

surement location. A vertical array of six 4-MHz

UDVP (Ultrasonic Doppler Velocity Profiler) probes,

positioned at 6, 16, 26, 36, 46 and 76 mm above the

bottom of the flume, recorded the streamwise compo-

nent of flow velocity upstream of the probes, follow-

ing the technique described by Best et al. (2001). The

sampling rate of each UDVP probe was 4.5 Hz.

Located at the same height (except for 76 mm) and

adjacent to the UDVP probes were 5 siphoning tubes

of 6 mm diameter, which continuously sampled the

flow as it passed by. The suspension samples were

collected continuously in 5 aligned rows of 20 sample

containers, one row for each of the 5 siphoning tubes.

The 100 container array was set up on a sliding

trolley, which was rapidly advanced below the out-

flow ends of the siphon tubes at 4 s intervals, allowing

each successive column of 5 beakers to fill synchro-

nously over ~4 s. The tubes were 1.2 m long, and

b) particle size distribution.

Fig. 2. Frame captured from the video recording of the turbidity current of 14% initial concentration at Location 4, i.e., at 5.28 m from the inlet.

1, 2, 3, 4, 5 refers to siphon and UDVP probe positions, located at respectively 6, 16, 26, 36, 46 mm heights. 6 refers to UDVP probe only,

located at 76 mm height. The arrow points to the turbulent eddy, which generates a rapid deceleration event in the velocity time series. The field

of view is approximately 0.55 m long and 0.35 m high.


allowance was made for the measured transit times

when registering the timing of sample collection with

that of UDVP data collection The content of each

container was later analysed using a Malvern Master-

sizer Plus laser diffraction grain sizer, yielding grain

size distribution and suspended sediment concentra-

tion information at a rate of 0.25 Hz. We will show

below (Section 5.1) that the duration of the head,

delimitated by the horizontal velocity data, is 4 s for

the low-density experiments, i.e. the sampling dura-

tion, and about 2.5 s for the high-density experiments.

Care was taken that as little ambient fluid as possible

was collected prior to arrival of sediment-bearing fluid

at the siphon outlet. This implies that for the high-

density flows, the sampling beaker of the head will

have incorporated some body material for up to 1.5 s.

However, we will show, in Section 6.3, that the di-

mensionless concentration and grain size data in the

head and body are roughly similar; hence we assume

that any effect of dilution from the body material into

the head were insignificant.

A detailed study of the spatio–temporal evolution of

the downstream velocity, grain size distribution and

concentration of the flow was feasible. Moreover,

additional information on the spatio–temporal evolu-

tion of the downstream component of turbulence of the

flow was obtained from the calculation of root-mean-

square (RMS) values of downstream velocity. RMS

velocity is equal to the standard deviation of velocity

averaged over a certain time period (Kneller et al.,

1997; Buckee et al., 2001; Baas and Best, 2002).

Time series of RMS velocity were calculated by aver-

aging the instantaneous velocity data over a 2-s long

period along the whole duration of the velocity time

series and for each measurement height. Tests carried

out to verify the effect of other lengths of averaging

periods on the time series showed no significant differ-

ences. In order to remove the unwanted effect on RMS

velocity of long-term flow deceleration, the velocity

profile in each time window was de-trended using

standard linear regression analysis prior to calculating

the RMS velocities. In analysing velocity signals from

UDVP probes in still water, the long-term average

deviation from the mean was found to be between 2

and 3 mm s�1. These values are therefore considered a

threshold value for the UDVP instrument noise. Areas

of the experimental flows with RMS values below 3

mm s�1 need not be caused by turbulence.


4. Description of 14% experimental data

The time series of downstream velocity, grain size,

and concentration are presented for each sampling

location in Fig. 3. This figure also illustrates the

calculated root-mean-square (RMS) values of the

downstream velocity data (Fig. 3b). The reference

time is taken as the time from the removal of the

reservoir stopper. The flow arrival time ranges from

7 s at Location 1 to 35 s at Location 6.

4.1. Downstream horizontal velocity data

The six UDVP probes acquired time series of the

downstream horizontal velocity as the flow travelled

by. The temporal evolution of the streamwise velocity

field for each location is shown in Fig. 3a. A coherent

structure that evolves slightly between the different

locations is observed.

A zone in which the velocity values are consis-

tently high (up to 265 mm s�1) through the flow

depth is recorded for 2–3 s after the arrival of the

flow front (Fig. 3a). Thereafter the velocity drops at

each measurement height. The flow deceleration is

very rapid for the upper two probes (at heights 46

and 76 mm), forming a zone with velocities as low

as 40 mm s�1. The velocity decrease is reduced for

the lowermost probe (6 mm high), with velocities

reduced by a third of the maximum value at that

level. The probes at heights of 16, 26 and 36 mm

reveal a zone of high velocity, with the maximum

velocity occurring close to 20 mm. The height of

this interpolated maximum does not vary temporal-

ly, and varies only slightly spatially (+/�2 mm; Fig.

3a). However, the internal structure of this zone

does evolve with time and distance. At Location

1, the time series exhibits rapidly fluctuating flow

velocity for about 30 s after passage of the head.

These fluctuations become less conspicuous at in-

termediate locations and disappear distally (e.g.,

Location 5 in Fig. 3a). The flow duration, measured

as the time taken for the downstream velocity to

fall below a predefined reference value (cf., McCaf-

frey et al., 2003), progressively increases as the

flow propagates downstream. For example, using a

reference velocity of 100 mm s�1, duration

increases from 28 s at Location 1 to 32 s at

Location 5.

4.2. RMS downstream velocity data

The time series of the RMS of downstream

velocity is shown in Fig. 3b. All graphs exhibit

short-term changes in RMS velocity, which take

the form of small concentric structures in the con-

toured plots presented. The highest RMS values

were found immediately after flow arrival and for

a couple of seconds only, for the highest UDVP

probes. Rapid fluctuations in RMS values are also

found close to the bed at 6 mm. The periodicity of

these fluctuations is ~3 s. The time span between

the passage of successive zones of high RMS ve-

locity remains quasi-constant from Location 1 to

Location 5 whilst the RMS values decrease.

Above the velocity maximum, between 25 and 50

mm, and after the zone of highest RMS values, a

zone of intermediate RMS velocity exists (Fig. 3b).

RMS values are lowest at the level of the maximum

velocity and particularly during the last 10–15 s of

the time series. The most variable RMS time series

is observed at Location 1.

4.3. Grain size data

At all locations along the flow path, median

grain size evolves in a temporally and spatially

consistent way (Fig. 3c). At each location, and

during the first 5 s after the flow arrival, the

grain size data exhibit enrichment in coarse grains

relative to the initial particle distribution. Thereaf-

ter, the vertical grain size profile is characterised

by an upward decrease of the median grain size for

~25 s at Location 1, ~20 s at Location 2 and ~15

s at Location 3. Subsequently, the normal grading

changes into a characteristic vertical pattern of

normal to inverse grading for these locations. The

normal to inverse grading is particularly well-de-

veloped at Locations 4 and 5. At each location, the

zone of grain size reversal moves closer to the

base of the flow with time. A period during

which the flow carries relatively coarse grains is

seen near the base of the flow at about 10 s after

flow arrival at Location 1 and at 18–20 s after

flow arrival at the other locations (Fig. 3c). During

these periods, the flow contains the coarsest sedi-

ment measured, with D50-values of up to 7.7 Am(Location 1).

Fig. 3. Time series of a) downstream velocity (millimeter per second), b) calculated root-mean-square (RMS) values of downstream velocity (millimeter per second), c) median grain

size (micron), and d) concentration (volume percent), at six different flow heights 6, 16, 26, 36, 46 and 76 mm, for five different measurement locations, for the flows with 14% initial

concentration. Note that the 76 mm time series is only collected for downstream velocity. The time, in seconds, is expressed from the removal of the stopper from the bottom of the

reservoir, i.e. the time of inception of the flows. For each graph, the two dashed lines mark the position of the uppermost and lowermost probes, above which and below which no

more data are acquired. The scales and grey shades are the same as in Fig. 4, for easier comparison.

C.M

.A.Chouxet

al./Sedimentary

Geology179(2005)49–69

55

Fig. 4. Time series of a) downstream velocity (millimeter per second), b) calculated root-mean-square (RMS) values of downstream velocity (millimeter per second), c) median grain

size (micron), and d) concentration (volume percent), at six different flow heights 6, 16, 26, 36, 46 and 76 mm, for five different measurement locations, and for the flows with 5%

initial concentration. All data, except RMS velocities, were presented in McCaffrey et al. (2003), but have been redrawn at the same scale and grey shades as in Fig. 3. The time, in

seconds, is expressed from the removal of the stopper from the bottom of the reservoir, i.e. the time of inception of the flows. For each graph, the two dashed lines mark the position of

the upper and lower probes, above which and below which data extrapolation is likely to be biased.

C.M

.A.Chouxet

al./Sedimentary

Geology179(2005)49–69

56


4.4. Concentration data

The time series of sediment concentration as a

function of flow height are given in Fig. 3d. Concen-

tration consistently decreases upwards. At each mea-

surement location, the concentration values are

initially very low (down to 2–3 vol.%) compared

with the initial concentration of 14 vol.% within the

reservoir. The maximum concentration, located at the

base of the flows, occurs at each location some 15–20

s after passage of the flow front. The maximum

concentration at Location 1, which is closest to the

inlet, equals 6.4%. At other locations, the maximum

measured concentration ranges from 5.9% at Location

2 to 7% at Location 5, i.e., furthest away from the

inlet. The interpolated near-bed concentrations values

are 6.4% at Location 2 and 7.9% at Location 5. The

maximum heights reached by concentration contours

V4.5% gradually decrease in a downstream direction.

5. Interpretation of high-density flow data and

comparison with low-density flow data

In this section, the data of the high-density experi-

ments are interpreted and compared with the low-

density experiments of McCaffrey et al. (2003). The

spatio–temporal graphs of the low-density experi-

ments are reproduced in Fig. 4 to facilitate the com-

parison. Fig. 4 also includes time series for RMS

values of downstream velocity, which have not been

published before. Due to the fact that the low-density

flows were slower than the high-density flows, the

low-density flows arrived later at each measurement

location than the high-density flows, explaining the

different initial times on the graphs in Figs. 3 and 4.

Particulate gravity currents are commonly divided

into three flow regions: head, body and tail (see

review by Kneller and Buckee, 2000). The head,

with its overhanging nose due to no-slip condition at

the lower boundary and frictional resistance at the

upper boundary, is the area where mixing of the

current with ambient fluid occurs, essentially by

detraining of dense fluid out of the back of the head

in a series of transverse vortices (Allen, 1971; Britter

and Simpson, 1978; Simpson and Britter, 1979). The

body is the area which has a thin, relatively dense

layer of fluid near the base of the current, and which is

overlain by a mixing zone at its upper boundary

displaying succession of large eddies (Ellison and

Turner, 1959; Middleton, 1966). The tail is the termi-

nal part of the flow, where velocity is low and grad-

ually decreases to zero; here slow settling from

suspension is the dominant depositional process.

The same subdivision is applied below, because it

was possible to confidently delimit head, body and

tail by trends in the experimental data (see also

McCaffrey et al., 2003).

5.1. Downstream velocity

The head of the flow with 14% initial concentra-

tion is delimited by the rapid increase in velocity at

the flow front (Fig. 3a) and the midpoint of a slightly

longer period of relatively strong flow deceleration

present at heights of 46 and 76 mm in all locations.

The head passes the measurement locations in about

2–3 s. The strong deceleration may relate to the

presence of an eddy at the back of the head; analysis

of video recordings of the experiments confirms the

existence of such a structure, located at the back of,

and defining the extension of the head (Fig. 4). A

more gradual flow deceleration event at 20–25 s,

prominent in particular at proximal locations, defines

the transition between the body and tail of the flow. It

corresponds to the beginning of the waning tail of the

flow, probably because the reservoir empties and thus

no longer supplies the flow (the reservoir was seen to

empty at ~21.5 s after the start of the experiments).

The velocity structure of the high-density flow

evolves in time and with distance along the flume.

As a result of the absence of a bed slope and progres-

sive deposition of sediment, flow velocity gradually

decreases at all levels in the flow and at all locations

along the flume. Also, the short-term fluctuations in

flow velocity, which were particularly clear at Loca-

tion 1, gradually disappear. The increasing duration of

the flows away from the inlet is primarily caused by

extension of the tail away from the inlet, because the

duration of the passage of the head and body is almost

constant at the studied locations.

In general terms, the high- and low-density flows

have similar time-dependent velocity structure. In

detail, however, there are important differences. The

dense flow is thinner and travels faster than the dilute

flow, with a shorter head duration (about 2.5 s instead


of 4 s for the low-density flows). The height of

maximum velocity is lower for the high-density

flow (c20 mm) than for the low-density flow

(c25 mm). However, the velocity maximum is

located roughly at about 0.3 of the height of the

head in each case, which is in agreement with

previous experiments (Altinakar et al., 1996; Kneller

et al., 1997, 1999; Best et al., 2001).

5.2. RMS downstream velocity

The time series of RMS downstream velocity,

shown in Figs. 3b and 4b, indicates that, overall, the

flows of 14% initial concentration have higher RMS

values than the flows of 5% initial concentration,

although the distribution of RMS values is similar.

Below, the RMS velocity structure of the low-density

flows is described first, then the RMS velocities of the

low- and high-density flows are compared.

All the graphs of RMS velocity for the low-density

flows exhibit numerous rapid fluctuations in the head

of the flows as well as close to their base (Fig. 4b).

The highest RMS values are obtained from the upper

front of the head at Location 2. Locations 3–5 have

their maximum RMS velocities at the same position,

but absolute values decrease downstream. Location 1

is characterised by strong fluctuations in RMS veloc-

ity in the entire head and body. The rear part of the

flows and the upper part of their bodies are charac-

terised by low RMS velocities (Fig. 4b). Particularly

striking are the regular fluctuations in RMS values

near the base of the flow. As for the high-density

flows, a zone of relatively low RMS values exists at

the level of the velocity maximum, above which

intermediate fluctuating RMS values are observed.

In both the low- and high-density flows, the zone

of maximum RMS velocity within the upper part of

the head is interpreted to result from high turbulence

levels linked to the friction between the propagating

flow and the ambient fluid, leading to the formation of

Kelvin–Helmholtz waves and mixing at the back of

the head (Best et al., 2001). The concentric RMS

structures near the base of the flows, whose periodic-

ity slightly increases with time, are interpreted as

coherent flow structures (Baas et al., in press),

corresponding to turbulent eddies generated by fric-

tion at the lower flow boundary. The relatively low

RMS velocities at the level of maximum velocity,

supporting previous measurement by Kneller et al.

(1999), Best et al. (2001) and Buckee et al. (2001)

as well as in the tail of the flows indicate that these

areas are less turbulent than other areas. It thus

appears that the propagation distance of turbulent

eddies generated by shear at the lower, upper and

frontal flow boundaries is relatively small (cf. Felix

et al., 2005). The degree of turbulence decreases

downstream along the flume in both flows, which

correlates with decreasing downstream velocity and

thus decreasing shear.

5.3. Median grain size

The median grain size structures of the high-den-

sity flows (Fig. 3c) and the low-density flows (Fig. 4c)

evolve in a similar way. Yet, the basal zone of maxi-

mum grain size is less well developed in the low-

density flows. At each location, the grain size helps to

define the transition between the head and the body of

the flow. A drop in median grain size by up to 1 Ammarks this transition.

The inverse grading in the body of the flows is

interpreted to result from the movement of coarse

sediment from the upper part of the head, enriched

in coarse particles (cf., Section 4.3), upwards and

backwards by turbulent motion (McCaffrey et al.,

2003). During the backward motion, the coarse sedi-

ment is probably located above the measurement area,

i.e. above the highest sampling tube located at 46 mm.

They then fall passively back into the body and tail of

the flow. Video data (Fig. 2) and velocity (Figs. 3a and

4a) support the interpretation that an eddy is present at

the back of the head, which may be responsible for

this redistribution of coarse sediment towards the rear

of the flows.

A major difference between the low- and high-

density flows is that a basal zone enriched in coarse

grains is present at 18–20 s after the passage of the

head at Locations 2–5 in the high-density experiments

(Fig. 3c). This phenomenon was observed only at

Locations 3 and 5 in the low-density experiments.

Their position below the velocity maximum and far

behind the head classifies these coarse-grained zones

as coarse tail lags sensu Hand (1997). The fastest

settling grains are thus concentrated towards the

base of the flow, while slower settling grains are

distributed more evenly throughout the flow depth

Fig. 5. Spatio–temporal evolution of a single idealised flow, created by using the data acquired at the five measurement locations, for a) downstream horizontal velocity (millimeter

per second), b) RMS velocity (millimeter per second), c) median grain size (micron), and d) concentration (volume percent), at 21.5, 29, 36.5, 44 and 53.5 s after inception of the

flows. The time, in seconds, is expressed from the removal of the stopper from the bottom of the reservoir, i.e. the time of inception of the flows. The grey dashed lines show the

position of the uppermost and lowermost probes respectively above which and below which no more data are acquired. A small contouring artefact is noticeable before the flow

arrival, mainly visible for the downstream horizontal velocity and RMS velocity graphs, at 21.5 and 29 s.

C.M

.A.Chouxet

al./Sedimentary

Geology179(2005)49–69

59


(Middleton and Southard, 1984; Hand, 1997 and

references therein).

5.4. Concentration

The initial sediment concentration in the reservoir

was 14%. The fact that a maximum concentration of

6.4% was observed at Location 1 (Fig. 3d) indicates

that strong flow dilution occurred due to flow expan-

sion and entrainment of ambient water and/or sedi-

mentation. In the flows of 5% initial concentration,

McCaffrey et al. (2003) also noticed an abrupt change

in concentration between Locations 1 and 2 (Fig. 4d).

They explained the reduction of nearly 50% by in-

voking high rates of sedimentation from suspension

between those locations. In the high-density experi-

ments (Fig. 3d), no such drastic change is observed.

This point will be discussed in more detail in the

section on dimensionless analysis below. The progres-

sive spatio–temporal decrease in the height of con-

centration contours, from Locations 1 to 5, attests to

ongoing sedimentation and dilution as the flow pro-

pagates, although these processes are less marked than

for the low-density flows. At Location 5, the maxi-

mum measured concentration of 7% by volume, seen

at the base of the flow at around 55 s, is higher than

the corresponding maximum seen in Location 4,

immediately upstream, at around 45 s, which is

equal to 6% by volume, interrupting the overall pat-

tern of downstream-decreasing concentration. A pos-

sible explanation is that the flow undergoes a slight

increase in the rate of fallout of sediment from sus-

pension between Locations 4 and 5 due to decreasing

velocity and turbulence intensity.

5.5. Spatial flow evolution

A series of instantaneous snapshots of the high-

density flow was constructed at five selected times

(i.e., 21.5, 29, 36.5, 44 and 53.5 s), for downstream

velocity (Fig. 5a), RMS velocity (Fig. 5b), median

grain size (Fig. 5c) and concentration (Fig. 5d). This

was done by extracting the respective data for each

point in time and for each measurement location from

the time series, and then plotting the data as a function

of distance along the flume for each point in time.

These spatial plots permit the visualisation of the

internal structure of a single flow and its temporal

evolution. They would also allow for a direct com-

parison with numerical modelling results, as sug-

gested by Felix (2002).

The snapshots of downstream velocity (Fig. 5a)

show maximum values at the front of the head and

progressively slower flow with increasing distance

behind it. The decrease in velocity in the tail part is

particularly evident around the height of maximum

velocity. The corresponding snapshots of RMS veloc-

ity data (Fig. 5b) do not show a steady evolution.

Generally, the head of the flow has the highest RMS

values, and therefore is the most turbulent part of the

flow. In the body, RMS velocities are clearly less than

in the head, except for the basal part of the flow,

where turbulence remains strong even at large dis-

tances behind the flow front. The spatial plots of

median grain size (Fig. 5c) show that the zone of

minimum grain size becomes more pronounced as

the flow evolves temporally and that coarse tail lag-

ging occurs behind the head of the flow (e.g., at 44s).

The spatial plots of concentration reveal a higher rate

of sedimentation as time goes by. Indeed at 29 s, the

height between the 2% and 6% concentration contours

is about 40 mm whereas it is only 20 mm at 53.5 s. A

zone of maximum concentration is visible close to the

base of the flow (Fig. 5d). It moves progressively

down the flume as the flow evolves. The zone is

centred at ~1 m after 29 s, at 2.5 m after 36.5 s, at

6.5 m after 44 s and beyond 7 m after 53.5 s.

The temporal evolution of the internal structure of

the high-density flow seen in Fig. 5 for all measured

flow and sediment parameters, indicates that the flow

is non-uniform. Decreasing velocities and overall con-

centrations indicate that this is probably caused by the

flow’s depositional character. This interpretation rein-

forces the conclusions drawn by McCaffrey et al.

(2003) and extends them to flows of higher concen-

tration. The fact that the low- and high-density flows

are non-uniform implies that it is impossible to deduce

the structure of flows from studies of time series data

obtained at only one location, at least over the concen-

tration range encompassed by the flows as they evolve.

6. Dimensionless analysis

Dimensionless analysis was carried out in order to

compare the flow structure of the low- and high-


density flows in detail, and to investigate if data

collapse could be achieved. Here, the flows are com-

pared by means of dimensionless vertical profiles of

normalised downstream velocity (Fig. 6), RMS veloc-

ity (Fig. 7), median grain size (Fig. 8) and concentra-

tion (Fig. 9). A careful choice of normalisation

parameters is essential in order to ensure that flow

and sediment parameters are compared in analogous

zones of the flows (cf., Felix, 2004). Therefore, the

dimensionless parameters used in this study are de-

fined first. Thereafter, the location of vertical profiles

in the head, body and tail of the turbidity currents are

selected. The dimensionless profiles of flow and sedi-

ment parameters are presented in the last part of this

section.

6.1. Normalisation parameters

Selection of reference heights, velocities, median

grain sizes and concentrations was necessary in order

Fig. 6. Dimensionless flow velocity as a function of dimensionless height a

low- and high-density (5% and 14% initial concentration, respectively) tu

to normalise and compare the data from the two sets

of experiments. The height of the maximum down-

stream velocity was used as reference for the calcula-

tion of normalised height (cf., Altinakar et al., 1996;

Kneller et al., 1999). This reference height, which was

shown to be independent of measurement location, is

~25 mm for the low-density flows (Fig. 4a) and ~20

mm for the high-density flows (Fig. 3a). Other refer-

ence heights, such as the height at which the down-

stream velocity in the upper part of the flow is half the

maximum velocity (e.g., Kneller and Buckee, 2000),

could not be reliably applied because there were too

few sampling heights above the velocity maximum to

accurately determine them.

In the vertical profiles for the head (Fig. 6a) and

body (Fig. 6b), dimensionless downstream flow ve-

locity was defined as the ratio between average ve-

locity over a time span of 1.25 s (flow of 14% initial

concentration) or 2 s (flow of 5% initial concentra-

tion) and average head velocity for all measurement

nd measurement location for the head (a), body (b) and tail (c) of the

rbidity currents.

Fig. 7. Dimensionless RMS velocity as a function of dimensionless height and measurement location for the head (a), body (b) and tail (c) of the

low- and high-density (5% and 14% initial concentration, respectively) turbidity currents.


locations. The time spans of 1.25 and 2 s were chosen

to obtain velocity values over approximately equal

lengths in both flows. Average downstream head ve-

locities, at the height of the velocity maximum, were

133.5 and 189 mm s�1 for the low- and high-density

flows, respectively. The use of a single head velocity

for each flow concentration is warranted, because

head velocity changes between the most proximal

and most distal measurement locations were insignifi-

cant. The calculation of the dimensionless flow veloc-

ity for the tail zone (Fig. 6c) was carried out using the

maximum velocity found at the body–tail transition

because the head and tail are sufficiently far apart that

processes in the head may not be relevant to tail

dynamics.

The dimensionless RMS velocity for the head (Fig.

7a) and body (Fig. 7b) was calculated by normalising

the RMS velocity values to the average head velocity.

For the tail (Fig. 7c), the same method was used as for

the normalised downstream velocity, i.e., the maxi-

mum downstream value measured at the body–tail

transition was selected.

Median grain size (Fig. 8) was normalised to the

initial median grain size (8 Am) in the overhead

reservoir for the low- and high-density flows. Dimen-

sionless concentrations (Fig. 9) were calculated by

dividing the local concentration values by the initial

concentration in the overhead reservoir.

6.2. Division of flows into head, body and tail

segments

Vertical profiles of flow parameters were outlined

through predefined segments in the experimental

flows. The underlying methodology relies on the de-

termination of equivalent zones in the flows in each of

the two sets of experiments, and subsequent selection

of equivalent locations within these zones where ver-

tical profiles are to be compared. Below, the zones are

defined in terms of the flows’ head, body and tail.

Fig. 8. Dimensionless median grain size as a function of dimensionless height and measurement location for the head (a), body (b) and tail (c) of

the low- and high-density (5% and 14% initial concentration, respectively) turbidity currents.


6.2.1. Head length

In the time series, the passage of the head is

defined as the period immediately following the ar-

rival of the flow, during which the velocities are high

along the entire vertical profile (Fig. 3a). The up-

stream boundary of the head, i.e. the transition be-

tween head and body, is defined by the rapid

decelerating event in the velocity time series, recorded

by the probes at 46 and 76 mm height. The average

period in which the head passed a measurement loca-

tion was 4 s in the low-density flow and 2.5 s in the

high-density flow. In length, this corresponds to about

0.38F0.005 m for both flows, using the average head

velocity as reference. It was then decided arbitrarily to

select a relative distance of 25% of the head length

behind the front of the head to locate the vertical

profiles. Thus, at 0.095 m from the front of the

head, the original data located along a vertical profile

were chosen for normalisation. Because of the large

fluctuations of the downstream and RMS velocity,

instead of presenting an isolated profile from this

location, average velocity values were calculated for

all the values between 0 and 0.19 m (giving the

average velocity value at 0.095 m).

6.2.2. Body length

The body stretches from the upstream limit of the

head to the point marked by a sudden change from

high to low velocity (at proximal locations, Fig. 3a)

and a change from relatively high to low RMS veloc-

ity (predominantly at distal locations, Fig. 3b). This is

interpreted to represent the time when the overhead

reservoir emptied. The reservoir emptied in 22 s for

the low-density flow and 21.5 s for the high-density

flow. Allowing for the head duration, the duration of

the passage of the flow bodies in the low- and high-

density flows was thus 18 and 19 s, respectively, with

corresponding respective body lengths of 1.73 and

Fig. 9. Dimensionless concentration as a function of dimensionless height and measurement location for the head (a), body (b) and tail (c) of the

low- and high-density (5% and 14% initial concentration, respectively) turbidity currents.


2.85 m (once again using average head velocity as the

reference velocity). The vertical profiles for dimen-

sionless downstream velocity and its RMS values,

median grain size and concentration were drawn at

an arbitrary relative distance of 30% of the length of

the body from its front, hence at 9.4 s and 0.90 m for

the low-density flows, and 8.2 s and 1.23 m for the

high-density flows. Here, 9.4 and 8.2 s refer to the

time since the arrival of the flow.

6.2.3. Tail length

The tail is the most distal part of the flows bounded

by the body at one end. In the experiments, the tail

was disrupted by the arrival of flow reflections before

it had come to rest. Because the tail is the section

where most of the flow stretching takes place, a

different method is required to select an equivalent

location for the vertical profiles. First, a power func-

tion was fitted to the velocity time series at a height of

2.6 cm (i.e., close to the level of maximum velocity)

in the tail of each flow. Subsequently, the best fit

power function was used to calculate the time period

from the time of first arrival of the tail to the time at

which velocity reached 10 mm s�1. Finally, the time

for the tail velocity to decrease by an arbitrary 40% of

the range between its value at the body–tail boundary

and the 10 mm s�1 boundary was calculated for each

measurement location. At these times vertical profiles

for downstream flow velocity and RMS velocity,

concentration and median grain size were determined.

As mentioned above, dimensionless velocities were

calculated by dividing the values at the 40% boundary

by the velocity at the body–tail boundary rather than

by the average head velocity, because the head and tail

are so far apart that head processes should not affect

the tail.

6.3. Dimensionless vertical profiles

The normalised downstream velocity (Fig. 6),

RMS velocity (Fig. 7), median grain size (Fig. 8)

and concentration (Fig. 9) are plotted versus the di-

mensionless height for each of the three zones of the

flow (i.e., body, head and tail).


6.3.1. Downstream velocity profiles

For the head and body (Fig. 6a and b), normalised

downstream flow velocities greater than 1 indicate

flow towards the front of the turbidity current, while

normalised velocities smaller than 1 signify flow that

moves away from the head for an observer moving

with the flow (i.e., within a Lagrangian reference

frame). For the tail (Fig. 6c), velocities are relative

to that of the body–tail transition. The normalised

downstream velocities for the low- and high-density

flows at Locations 2–5 are similar for the head (Fig.

6a), body (Fig. 6b) and tail (Fig. 6c), hence the data

collapse in a satisfactory manner. At Location 1,

however, the normalised velocity of the high-density

flow is significantly higher than that of the low-den-

sity flow (Fig. 6a), particularly within the head. Be-

tween Locations 1 and 2 the dimensionless values for

the low-density experiment increase drastically, thus

supporting the inferred occurrence of an episode of

high sedimentation rate and change from jet to tur-

bidity current (McCaffrey et al., 2003) between these

locations. In contrast, the dimensionless velocity

remains quasi-constant between Locations 1 and 5

in the high-density flow. This suggests that no regime

change occurred along this transect. The flow may

therefore have developed into a turbidity current by

the time that it reached Location 1. In turn, this

implies that any episode of high sedimentation rate

must have occurred upstream of Location 1, and thus

within 1.32 m of the inlet.

6.3.2. RMS downstream velocity

The vertical profiles of dimensionless RMS ve-

locity versus height (Fig. 7) in the head region of

the flows (Fig. 7a) are irregular, with little similarity

between the low- and high-density experiments.

However, a broad trend with higher RMS velocity

values at the top of the flow compared with the rest

of the profile and a slight increase of RMS close to

the base of the flow, exists. In the body (Fig. 7b),

normalised RMS values are more uniform than in

the head, and maximum RMS velocities are almost

exclusively found at the base of the flows. A weak

zone of relatively low RMS velocities is discernable

at or around the height of the velocity maximum,

particularly at distal locations. The RMS velocities

in the tail of the low- and high-density flows col-

lapse well (Fig. 7c), displaying a quasi-uniform

pattern of RMS velocities along the entire flow

depth, except for a slight increase at the lowest

data point.

At Location 1, the vertical profile of the flow of 14

initial concentration is broadly concave to the right (as

are the profiles at all other Locations), whereas the

profile of the flow of 55 initial concentration is con-

cave to the left. This difference is interpreted to arise

because the high-density flow has undergone the jet to

turbidity current transition at this Location, whereas

the low-density flow still has elements of jet structure.

The higher dimensionless RMS values observed near

the base of the body and tail profiles represent the

turbulent eddies generated by friction with the base of

the flume. This pattern is also seen close to the base of

the flow for the head, yet the basal non-dimensional

RMS values are less than those in the upper part of the

flow, confirming that the head is more turbulent at its

top than at its base. The profiles also show the general

loss of turbulence as the flow propagates. The profiles

become quasi-vertical straight lines with RMS values

close to zero in the tail areas, which indicate that the

flow approaches a laminar regime over its entire

height.

6.3.3. Median grain size profiles

The vertical profiles of median grain size (Fig. 8)

essentially redisplay the data in Figs. 3c and 4c, but

now allow a more direct comparison between the low-

and high-density flows. The vertical profiles of the

low- and high-density flows generally exhibit a simi-

lar trend for the head (Fig. 8a), body (Fig. 8b) and tail

(Fig. 8c). At most locations, the sediment is relatively

coarse near the base of the flow. In the central part of

the flows, the sediment is relatively fine, while its size

increases slightly in the uppermost part of most flows.

The height of minimum grain size decreases from

body to tail and from Locations 2 to 5, the grain

size minimum being more prominent distally than

proximally. Location 1 once again differs from the

other locations in that the normal-to-inverse grading

in the low- and high-density flows cover different

dimensionless size ranges, and normally graded pro-

files (high-density flow) versus weakly graded to non-

graded profiles (low-density flow) prevail in the body

and tail. Once again, these discrepancies suggest that

the flows may not have developed to the same state at

Location 1.


6.3.4. Concentration profiles

The vertical profiles of dimensionless concentra-

tion are presented in Fig. 9. The vertical profiles show

a good collapse of concentration data, especially in

the body (Fig. 9b) and tail (Fig. 9c) of the flows. The

exception is again Location 1, where dimensionless

concentration for the low-density flow is consistently

higher than for the high-density flow. In all the graphs

of Fig. 9, the highest normalised concentration is

observed at the base of the flow.

Fig. 9 reinforces the above-mentioned conclusion

that the low- and high-density flows had not evolved

to the same state when they reached Location 1. The

flows show similar nondimensional behaviour from

Locations 2 to 5. The high-density flow is compara-

tively more dilute than the low-density flow when it

reaches Location 1, having lost most of its particulate

load by sedimentation while the flow was sedimenting

rapidly. We infer that the high-density flow underwent

a period of strong sedimentation and change from

inbound jet to turbidity current, before reaching the

most proximal measurement location, at which point

the low-density flow was still propagating with a

relatively high concentration.

7. Discussion and conclusions

Detailed experiments with flows of 14% initial

concentration were undertaken to gain insights into

the spatio–temporal development of flow structure in

terms of horizontal velocity, RMS velocity, concen-

tration and grain size, and to make a comparison

with the evolution of flows of 5% initial concentra-

tion, as described by McCaffrey et al. (2003). The

high-density flows were shown to be non-uniform.

The current duration extends via the stretching of the

tail only, so that the length and duration of the head

and body remain constant throughout the flow length

studied. Behind the head of the flow and below the

maximum velocity zone, a coarse tail lag in sus-

pended sediment has been observed at all locations.

These experimental data therefore strongly support

the model of Hand (1997), in which coarser grains

lag behind the head of the current by virtue of being

concentrated at a level below that of the velocity

maximum, therefore advecting forward more slowly

than the relatively finer-grained material carried

higher up. Although the flows of 14% initial con-

centration are non-uniform, the dimensionless analy-

sis shows a generally good collapse of the data from

the low- and high-density flows for the different

parameters studied. It follows that the fundamental

flow structure is essentially the same for flows of

either initial starting concentration at any stage in

their development after the jet to turbidity current

transition. The implications of this conclusion are

considered below.

The common evolution of the flows of 5% and

14% initial concentration might suggest that the par-

tial turbulence suppression predicted to develop at

concentrations above 9% (Bagnold, 1962; Shanmu-

gam, 2000; Gani, 2004) does not appear to influence

the development of the high-density flow. It should be

borne in mind, however, that flows of both initial

concentrations were considerably more dilute at Lo-

cation 1, i.e. 1.32 m from the flow inlet, than when

generated. Such a dilution could be produced by

turbulent entrainment of water and/or by sedimenta-

tion. McCaffrey et al. (2003) sampled two flows (their

Flows 1 and 2) with one siphon tube/UDVP probe

pair located 0.04 m from the outlet of the reservoir,

and found that concentration had fallen from 5% to

2.5% over this short distance. In the absence of a

significant deposit at this location, it was concluded

that rapid water entrainment had occurred. Thus, even

in the moderate Reynolds regime of the inbound jet

flow, high concentrations could only be maintained

transiently. This might imply that high-concentration

turbidity currents may not be viable as a long-range

transport mechanism. However, entrainment-related

dilution might also be an artefact of the flow genera-

tion mechanism. In these experiments the flows en-

tered the flume as a jet, necessarily incorporating

ambient water. Improved experimental design, or gen-

erating flows with yet higher initial concentrations,

might overcome this problem. Thus, the viability of

high density turbidity currents as a long-range sedi-

ment transport mechanism remains open. It seems

likely that no partial turbulence suppression was no-

ticeable because the turbidity currents produced did

not exceed the 9% concentration threshold of Bagnold

(1962). It follows that any dimensional analysis re-

garding flow evolution should not be applied to con-

centrations higher than those observed in these flows,

i.e. about 7%.


Even if the initial decrease of concentration is

dilution-related, it seems likely that sedimentation-

related concentration decreases may also have played

a role by Location 1. Certainly, such changes can be

inferred to have occurred between Locations 1 and 2.

Thus both low- and high-density flows may have

undergone deposition-related decreases in concentra-

tion, with the low-density flows reaching a lower

concentration than the high-density flows. This sug-

gests that the flows may have an upper concentration

threshold or capacity, indirectly controlled by the

inherited concentration, above which they cannot sus-

pend and transport the whole load of sediment. This is

presumably linked via the gravitational driving force

to its effect upon turbulence generation.

Finally, it is evident that after the jet to turbidity

current transition, the two data sets can be combined

to illustrate the path of an idealised flow of 14% initial

concentration (~7% concentration in the flume) until it

has developed the structure of a flow of 5% initial

concentration measured at the most distal location.

Such a flow would therefore develop over a length

scale greater than that provided by the experimental

facility. In addition, for new flows generated in similar

initial starting conditions but with initial concentra-

tions less than or equal to 14% by volume, dimen-

sionless concentrations could be used to calculate

local concentrations at various positions and times.

For these new flows, and as a first approximation,

head velocities could be computed by linear interpo-

lation, and then used (via the dimensionless values) to

estimate downstream velocities, concentrations and

grain size distributions of suspended material at var-

ious spatio–temporal positions.

Acknowledgements

We are grateful to T. Sakai, P. Julien and M. Felix

whose thorough reviews greatly improved and clari-

fied earlier versions of this manuscript. This research

was funded in part by the Turbidites Research Group

(TGR) Phase 4 programme, sponsored by BG, BHP-

Billiton, BP, ConocoPhillips, Shell and Norsk Hydro.

Caroline Choux acknowledges funding from Marie

Curie Individual Fellowship grant HPMF-CT-2001-

01298. The UDVP system was purchased under Natu-

ral Environment Research Council (NERC) Grant

GR3/10015 to Jim Best and colleagues. The overhead

reservoir flow generation mechanism and bslidingtrolleyQ siphon sample collection equipment were ini-

tially developed by J. Peakall. Complete data sets of

the experimental flows are available from the follow-

ing web site: http://trg.earth.leeds.ac.uk/.

References

Al-Ja’Aidi, O.S., McCaffrey, W.D., Kneller, B.C., 2004. Factors

influencing the deposit geometry of experimental turbidity cur-

rents: implications for sand-body architecture in confined

basins. In: Lomas, S.A., Joseph, P. (Eds.), Confined Turbidite

Systems. Geological Society, London, pp. 45–57.

Allen, J.R.L., 1971. Mixing at turbidity current heads and its

geological implications. Journal of Sedimentary Petrology 41,

97–113.

Allen, J.R.L., 1985. Principles of Physical Sedimentology. G. Allen

and Unwin, London. 271 pp.

Altinakar, S., Graf, W.H., Hopfinger, E.J., 1990. Weakly depositing

turbidity current on a small slope. Journal of Hydraulic Research

28, 55–80.

Altinakar, S., Graf, W.H., Hopfinger, E.J., 1996. Flow structure in

turbidity currents. Journal of Hydraulic Research 34, 713–718.

Baas, J.H., Best, J.L., 2002. Turbulence modulation in clay-rich

sediment-laden flows and some implications for sediment de-

position. Journal of Sedimentary Research 72, 336–340.

Baas, J.H., Van Kesteren, W., Postma, G., 2004. Deposits of deple-

tive high-density turbidity currents: a flume analogue of bed

geometry, structure and texture. Sedimentology 51, 1053–1088.

Baas, J.H., McCaffrey, W.D., Haughton, P.D.W., Choux, C.M.A., in

press. Coupling between suspended sediment distribution and

turbulence structure in a laboratory turbidity current. Journal of

Geophysical Research, Oceans.

Bagnold, R.A., 1962. Auto-suspension of transported sediment;

turbidity currents. Proceedings of the Royal Society of London.

Series A 265, 315–320.

Bagnold, R.A., 1966. The shearing and dilatation of dry sand and

the singing mechanism. Proceedings of the Royal Society of

London. Series A 295, 219–232.

Best, J.L., Kirkbride, A.D., Peakall, J., 2001. Mean flow and

turbulence structure of sediment laden gravity currents: new

insights using ultrasonic Doppler velocity profiling. In: McCaf-

frey, W.D., Kneller, B.C., Peakall, J. (Eds.), Particulate Gravity

Currents, Special Publication of the International Association

of Sedimentologists. Blackwell Science Ltd., Oxford, U.K.,

pp. 159–172.

Bonnecaze, R.T., Huppert, H.E., Lister, J.R., 1993. Particle driven

gravity currents. Journal of Fluid Mechanics 250, 339–369.

Britter, R.E., Simpson, J.E., 1978. Experiments on the dynamics

of a gravity current head. Journal of Fluid Mechanics 88,

223–240.

Buckee, C., Kneller, B., Peakall, J., 2001. Turbulence structure in

steady, solute-driven gravity currents. In: McCaffrey, W.D.,

http:\\trg.earth.leeds.ac.uk\



Special Publication of the International Association of Sedimen-

tologists. Blackwell Science Ltd., Oxford, U.K., pp. 173–187.

Choux, C.M.A., Druitt, T.H., 2002. Analogue study of particle

segregation in pyroclastic density currents with implications

for the emplacement mechanisms of large ignimbrites. Sedimen-

tology 49, 907–928.

Edwards, D.E., 1993. Turbidity Currents: Dynamics, Deposits and

Reversals. Lecture Notes in Earth Sciences, vol. 44. Springer-

Verlag, Berlin. 173 pp.

Ellison, T.H., Turner, J.S., 1959. Turbulent entrainment in stratified

flows. Journal of Fluid Mechanics 6, 423–448.

Felix, M., 2001. A two-dimensional numerical model for a turbidity

current. In: McCaffrey, W.D., Kneller, B.C., Peakall, J. (Eds.),

Particulate Gravity Currents, Special Publication of the Interna-

tional Association of Sedimentologists. Blackwell Science Ltd.,

Oxford, U.K., pp. 71–83.


49, 397–419.

Felix,M., 2004. The significance of single value variables in turbidity

currents. Journal of Hydraulic Engineering 42 (3), 323–330.

Felix, M., Sturton, S., Peakall, J., 2005. Combined measurements of

velocity and concentration in experimental turbidity currents.

Sedimentary Geology 179, 31–47.

Gani, R., 2004. From turbid to lucid: a straightforward approach to

sediment gravity flows and their deposits. The Sedimentary

Record, SEPM 2, 4–8.

Garcia, M.H., Parsons, J.D., 1996. Mixing at the front of gravity

currents. Dynamics of Atmospheres and Oceans 24, 197–205.

Gladstone, C., Phillips, J.C., Sparks, R.S.J., 1998. Experiments on

bidisperse, constant-volume gravity currents: propagation and

sediment deposition. Sedimentology 45, 833–843.


concentration, particle-driven gravity current. Physics of Fluids

10, 1083–1087.

Hand, B.M., 1997. Inverse grading resulting from coarse sediment

transport lag. Journal of Sedimentary Research 67, 124–129.

Hiscott, R.N., 1994. Traction carpet stratification in turbidites —

fact or fiction? Journal of Sedimentary Research A64, 204–208.

Kneller, B.C., Branney, M.J., 1995. Sustained high-density turbidity


ogy 42, 607–616.

Kneller, B.C., Buckee, C., 2000. The structure and fluid mechanics

of turbidity currents; a review of some recent studies and their


Kneller, B.C., Bennett, S.J., McCaffrey, W.D., 1997. Velocity and

turbulence structure of density currents and internal solitary

waves: potential sediment transport and the formation of wave

ripples in deep water. Sedimentary Geology 112, 235–250.

Kneller, B.C., Bennett, S.J., McCaffrey, W.D., 1999. Velocity struc-

ture, turbulence and fluid stresses in experimental gravity cur-

rents. Journal of Geophysical Research 104, 5381–5391.

Kuenen, P.H., 1966. Experimental turbidite lamination in a circular

flume. Journal of Geology 74, 523–545.

Laval, A., Cremer, M., Beghin, P., Ravenne, C., 1988. Density

surges: two-dimensional experiments. Sedimentology 35,

73–84.

Lowe, D.R., 1982. Sediment gravity flows; II Depositional models

with special reference to the deposit of high-density turbidity

currents. Journal of Sedimentary Petrology 52, 279–297.

Luthi, S., 1980. Some new aspects of two-dimensional turbidity

currents. Sedimentology 28, 97–105.

McCaffrey, W.D., Choux, C.M.A., Baas, J.H., Haughton, P.D.W.,

2003. Spatio–temporal evolution of velocity structure, concen-

tration and grain size stratification within experimental partic-

ulate gravity currents. Marine and Petroleum Geology 20,

851–860.

Mellor, G.L., Yamada, T., 1982. Development of a turbulent closure

model for geophysical fluid problems. Reviews of Geophysics

and Space Physics 20, 851–875.


rents. I Motion of the head. Canadian Journal of Earth Sciences

3, 523–546.


rents: III. Deposition of sediment. Canadian Journal of Earth

Sciences 4, 475–485.

Middleton, G.V., 1993. Sediment deposition from turbidity currents.

Annual Review of Earth and Planetary Sciences 21, 89–114.

Middleton, G.V., Hampton, M.A., 1973. Sediment gravity flows:

mechanism of flow deposition. In: Middleton, G.V., Bouma,

A.H. (Eds.), Turbidites and Deep Water Sedimentation: Society

of Economic Paleontologists and Mineralogists, Short Course

Notes, pp. 1–38.

Middleton, G.V., Hampton, M.A., 1976. Subaqueous sediment

transport and deposition by sediment gravity flows. In: Stanley,

D.J., Swift, D.J.P. (Eds.), Marine Sediment Transport and En-

vironment Management. Wiley, New-York, pp. 197–218.

Middleton, G.V., Neal, W.J., 1989. Experiments on the thickness of

beds deposited by turbidity currents. Journal of Sedimentology

Petrology 59, 297–307.

Middleton, G.V., Southard, J.B., 1984. Mechanics of Sediment

Movement, 2nd edition. Society of Economic Paleontologists

and Mineralogists. short course 3, 401 pp.

Mulder, T., Alexander, J., 2001. The physical character of subaque-

ous sedimentary density flows and their deposits. Sedimentolo-

gy 48, 269–299.

Mulder, T., Syvitski, J.P.M., Skene, K.I., 1998. Modelling of ero-

sion and deposition by turbidity currents generated at river

mouths. Journal of Sedimentary Research 68, 124–137.

Pantin, H.M., 1979. Interaction between velocity and effective

density in turbidity flow: phase–plane analysis, with criteria

for autosuspension. Marine Geology 31, 59–99.

Pantin, H.M., Leeder, M.R., 1987. Reverse flow in turbidity

currents: the role of internal solitons. Sedimentology 34,

1143–1155.

Peakall, J., Felix, M., McCaffrey, W.D., Kneller, B., 2001. Partic-

ulate gravity currents: perspectives. In: McCaffrey, W.D., Knel-

ler, B.C., Peakall, J. (Eds.), Particulate Gravity Currents, Special

Publication of the International Association of Sedimentologists.

Blackwell Science Ltd., Oxford, U.K., pp. 1–8.

Sanders, J.E., 1965. Primary sedimentary structures formed by

turbidity currents and related resedimentation mechanisms. In:

Middleton, G.V. (Ed.), Primary Sedimentary Structures and their

Hydrodynamic Interpretations — a Symposium, Society of


Economic Paleontologists and Mineralogists Special Publica-

tion, vol. 12, pp. 192–219.

Shanmugam, G., 1997. The Bouma sequence and the turbidite mind

set. Earth-Science Reviews 42, 201–229.

Shanmugam, G., 2000. 50 years of the turbidite paradigme

(1950s–1990s): deep-water processes and facies models —

a critical perspective. Marine and Petroleum Geology 17,

285–342.

Simpson, J.E., 1997. Gravity Currents in the Environment and the

Laboratory, 2nd edn. Cambridge University Press, New York.

Simpson, J.E., Britter, R.E., 1979. The dynamics of the head of a

gravity current advancing over a horizontal surface. Journal of

Fluid Mechanics 94, 477–495.

Stacey, M.W., Bowen, A.J., 1988. The vertical structure of density

and turbidity currents: theory and observations. Journal of Geo-

physical Research 93, 3528–3542.

Stix, J., 2001. Flow evolution of experimental gravity currents:

implications for pyroclastic flows at volcanoes. Journal of Ge-

ology 109, 381–398.

Stow, D.A.V., Reading, H.G., Collison, J.D., 1996. Deep seas.

In: Reading, H.G. (Ed.), Sedimentary Environments: Process-

es, Facies and Stratigraphy. Blackwell Science, Oxford,

pp. 395–453.

Zeng, J., Lowe, D.R., 1997. Numerical simulation of turbidity

current flow and sedimentation, I. Theory. Sedimentology 44,

67–84.



Morphology, evolution and fill: Implications for sand and

mud distribution in filling deep-water canyons and slope

channel complexes

Bryan T. Cronin a,*, Andrey M. Akhmetzhanov b, Adriano Mazzini a,

Grigorii Akhmanov c, Michael Ivanov c, Neil H. Kenyon b

TTR-10 Shipboard Scientists

aUniversity of Aberdeen, Department of Geology and Petroleum Geology, King’s College, Aberdeen AB24 2UE, United KingdombChallenger Division, Southampton Oceanography Centre, European Way, Southampton, SO14 3ZH, England, United Kingdom

cUNESCO Centre for Marine Geology and Geophysics, Science Park, Moscow State University, Russia


Abstract

A survey of the northeastern margin of the Rockall Trough on the Irish margin examined the transition from shelf edge to

basin floor, in morphology and sedimentary activity, of a deeply incised submarine canyon system, the Donegal Bay submarine

canyon. The survey produced superb 3D profiling of the canyon along its entire length, marking a transition from dcauliflowerTshaped head region with numerous tributary gullies feeding into one main canyon, to a single trunk canyon. This canyon, with

an initial combined width and depth of N17 km and N800 m in the dcauliflowerT head area, decreases rapidly to N4.5 km wide

and N450 m deep after the zone of tributary confluence. Eighteen kilometers further down dip, the canyon loses topographic

expression as it approaches the lower rise and floor of the Rockall Trough.

Degrees of recent sedimentary activity are evaluated by comparing side scan sonar systems of different frequency, and thus

of different penetration sub sea, and by ground-truthing using drop (gravity) cores. The canyon was a very active system,

dominated by sand transportation towards the floor of the Rockall Trough, along the slope as coarse-grained contourite, or as

sand spillover from the shelf. Sand was also deposited as overbank deposits outside the main head region of the canyon,

presumably by large volume turbidity currents and more active lateral gullies. The head area of the canyon system has been

progressively cut off from sand source by progressive sea level rise since the last glaciation. Sand was locally deposited on

terraces but not in the overbank area. Less frequent, lower volume and finer grained turbidity currents have become more

common in the system. The initial sand and bypass-dominated system with small sediment waves, which may be gravels, has

become dominated by muddy debrites in the lower reaches and by slumps in the upper reaches. Slumping in those upper reaches

leads to ponding of sand in the head and upper reach areas, with only very occasional turbidity currents transporting sand

further down the system in small channels.

0037-0738/$ - s

doi:10.1016/j.se

* Correspondi

E-mail addre

79 (2005) 71–97

ee front matter D 2005 Published by Elsevier B.V.

dgeo.2005.04.013

ng author.

ss: [email protected] (B.T. Cronin).

B.T. Cronin et al. / Sedimentary Geology 179 (2005) 71–9772

A model is produced to explain the mechanism and expression of backfilling in a large deep-water canyon system whose

hinterland has been flooded back since glacial drawdown of sea level in the eastern Rockall Trough area. This model explains

how sand may be trapped in large volumes in the upper reaches of a canyon system, due to slumping from the canyon margins

and nearby upper slope regions. The focusing of sand deposition in areas where this is not usually expected will have important

implications for hydrocarbon explorationists who wish to map the distribution of potential reservoir sand bodies within large,

confined deep-water canyon systems.

D 2005 Published by Elsevier B.V.

Keywords: Deep-water canyon; Rockall margin; Debrites; Backfilling; Slumps; Sand ponding

1. Introduction

1.1. Canyon morphology

There are few documented examples of canyons

from their head regions down to where they usually

lose topographic expression on the lower slope,

though some recent papers illustrate seafloor render-

ings from 3D seismic datasets (e.g. Barrufini et al.,

2000). Industrial seismic datasets are rarely collected

on lower slope areas where working petroleum sys-

tems are assumed not to work due to the lack of burial

of potential source rocks, so most of these renderings,

though usually spectacular, do not extend the full

length of the canyon system. There are some exam-

ples of sidescan sonar datasets where canyons have

been followed for long distances (Cronin et al., 1995).

For the moment, we still rely on conventional models

for deep-water canyons (e.g. Shepard, 1977; Scruton

and Talwani, 1982).

As a general rule, canyons typically have sinu-

ous courses with straight sections; floors which

deepen seawards; a V-shaped profile which is lost

as the adjacent continental slope reduces in gradi-

ent; tributaries, which are normally gullies in the

canyon head region; and currents which move up

and down their axes even at very great depth,

typically with tidal periodicities, or episodic turbid-

ity currents. They are normally associated with

large rivers (e.g. Amazon Canyon: Heezen and

Tharp, 1961; Damuth et al., 1983); rare on gentle

slopes, and densely spaced on steeper slopes; ero-

sive into any substrate; and older features, where

the canyon head is more deeply recessed into the

slope, or indeed, back across the shelf, indicating

that canyons develop shorewards by headward ero-

sion (Shepard and Marshall, 1978). Younger can-

yons are typically incised several hundred meters

below the shelf break. The morphologies of can-

yons are usually divided into four geomorphological

categories (Cronin, 1994):

(i) Canyon heads: Canyon head regions have been

described as amphitheatre-shaped areas with

steep rims (Belderson and Stride, 1969; Kenyon

et al., 1978). The oldest canyon on any one

slope is usually the canyon whose head is high-

est on the continental slope. Though tidal peri-

odicities are recorded in the lower reaches of

canyons using current meters, turbidity currents

are thought to be the most important process in

the head region (Shepard and Marshall, 1978;

Shepard, 1982). In summary, conventional mod-

els for canyon heads indicate that they are active

features which erode upslope and will ulti-

mately erode across the shelf; they entrain

shelf sediments, and funnel them basinwards;

they usually have currents moving through them

most of the time, both up and down canyon

axis; and turbidity currents account for most

of the intermittent flow through canyon heads,

and in combination with slumps, slides and

debris flows, are thought to be the main cause

of headward erosion.

(ii) Canyon axes: Canyon axes may be straight for

some sections but are locally sinuous, and are

commonly deflected by fault line escarpments

(e.g. Cap Ferret, Cap Breton and Guinivec Can-

yons, Brittany: Kenyon et al., 1978; Baltimore

and Wilmington Canyons, northeastern US con-

tinental slope: Twichell and Roberts, 1982;

Almeria Canyon, SE Spain: Cronin, 1995, Cro-

nin et al., 1995), or other seafloor topography

such as slumps (e.g. Lagos Canyon, Gulf of

B.T. Cronin et al. / Sedimentary Geology 179 (2005) 71–97 73

Cadiz: Gardner and Kidd, 1983; Monterey Fan

Canyon: Gardner et al., 1991) or salt domes

(Shepard and Emery, 1973). Otherwise, the

chief control on planform geometry of deep-

water canyons appears to be slope gradient

(Belderson and Kenyon, 1976). In summary,

deep-water canyon axis orientation is usually

driven by underlying structure, courses are

strongly affected by a range of seafloor topog-

raphy, and spacing appears to be controlled by

slope gradient.

(iii) Canyon walls: Canyon walls are usually V-

shaped, and this profile changes downslope,

in contrast to the U-shaped gullies which feed

the canyons in the head region, and of deep-

sea channels, which are usually flat-bottomed

(Carter, 1988). Most canyon walls are steep

or vertical, with local terracing giving most

canyons a dsteerheadT profile (e.g. Gollum

Channel System: Wheeler et al., 2003). A

downslope change in canyon profile from

V-shaped initially to U-shaped until the can-

yon eventually loses topographic expression,

has been observed by many workers (e.g.

Malahoff et al., 1982; Twichell and Roberts,

1982).

(iv) Distal canyon reaches: Distal parts of canyons

have not been imaged routinely. Most documen-

ted observations of lower canyon reaches date

back to the 80s. Canyons develop thalwegs in

their lower reaches and these thalwegs channels

may have significant dimensions (e.g. Stoe-

chades Canyon thalwegs, 300–500 m wide

and 25–75 m deep: Le Pichon and Renard,

1982; St Tropez Canyon, 200 m wide and 700

m deep: Le Pichon and Renard, 1982). These

thalwegs are also made remarkable by the fact

they rarely have a talus, which suggests that

lateral slumping of material from the canyon

and thalwegs walls is transferred into axial

transport within the thalwegs. These thalwegs,

found in most canyon lower reaches, pass into

constructional channels once the canyon has lost

topographic expression. There may be an inter-

vening transitional zone, called a canyon-fan

transition, which is characterized by a rise area

dissected by channels with an erosional aspect

(Akhmetzhanov et al., 2003).

1.2. Canyon evolution and filling

Current understanding of deep-water canyons, par-

ticularly those fed by subaerial drainage systems, is

that they form initially by retrogressive slumping on

the upper continental slope (e.g. Bouma et al., 1985:

Mississippi), and are maintained by funnelled density

flows. In studies of the Pliocene–Pleistocene eustatic

cycles in the Gulf of Mexico, the activity within the

Mississippi Canyon began as sea level fell (where

mass-transport complexes were deposited as the can-

yon was excavated). At the lowest point of sea level

sand was transported through the canyon into the deep

basin as channel levee systems, and as sea level began

to rise 30,000 years ago, deposition on the turbidite

fan ceased, and in the canyon area continued well into

the Holocene (12,000–11,000 years BP). As this pe-

riod corresponded to deglaciation, large volumes of

sediment continued to be funnelled through the Mis-

sissippi Canyon during transgression (Bouma et al.,

1989; Weimer, 1990). Sequence stratigraphic models

show cessation in turbidite deposition in shorter sub-

marine canyons, or those not fed by major rivers,

suggesting that submarine canyon activity stops just

prior to transgression (e.g. Posamentier and Vail,

1988).

Outcrop studies of canyon fills typically recognize

canyon fills that include chaotic deposits and thin-

bedded turbidites, with locally developed coarse-

grained, usually lenticular bodies, particularly in the

lower part of the fill (e.g. Doheny Channel, Piper and

Normark, 1971; San Carlos submarine canyon, Mor-

ris and Busby-Spera, 1988; Charo Canyon, Mutti et

al., 1988; Point Lobos submarine canyon, Clifton,

1981, 1984; Cronin, 1994; Cronin and Kidd, 1998).

In all of these examples, the submarine canyons are

interpreted to have been excavated initially due to

major external factors such as sea level fall or recon-

figuration of slope morphology by thrusting. Their

fills are then made up of one or more broadly fining-

upwards sequences, typically with clast-supported

conglomerates on an erosive surface (always inter-

preted as residual lags) and usually capped with thick

mud-prone intervals. In short, ancient submarine can-

yon fills are interpreted to reflect (a) erosion by

excavation of the upper slope by mass wasting; (b)

sediment bypass to the deep basin, (c) a major filling

phase; and (d) thin-bedded turbidites, probably con-


centrated by the local topography (Piper and Nor-

mark, 1971).

Clearly, a lot of what we know about the down-

slope evolution and interpreted behaviour of subma-

rine canyons on deep-water slopes is still rooted in

work from the 70s and 80s, with some excellent case

studies appearing in recent years. As more examples

of high resolution 3D seismic data showing renderings

of the seafloor, multibeam bathymetry surveys, and

deep-towed side scan sonar surveys of slopes are

acquired, we will develop a better understanding of

these systems. However at present, little is known

about the longer-term behaviour of canyons, and

their mechanisms of filling. This is important for oil

and gas exploration and production from slope can-

yons. Current thinking on the distribution of potential

reservoir facies, particularly sand bodies, has canyons

as zones of bypass as they develop, which backfill

with sand as their hinterland floods back, and then get

plugged with mud. This paper is an example of a

modern deep-water canyon that was surveyed from

the head area on the upper continental slope, down to

the upper rise area at the base of slope, with two

different side scan sonar tools. Ground-truthing of

the acoustic facies allowed the mapping of sand on

the present sea floor and in the shallow subsurface, to

test some of the models for sand distribution in these

features. It is hoped that the models produced will be

used as a proxy for predicting sand distribution in the

types of scenarios described.

The objective of this study was to examine the

planform expression of the Donegal Bay deep-water

canyon system on the NE Rockall Margin, and to

design a sampling strategy to recover sediment from

different parts of the canyon, to test the recent activity

of the system. Two side scan sonar systems were used,

the medium range OKEAN and the deep-towed high

frequency O.R.A.Tech systems, which operate at dif-

ferent frequencies and thus produce sonar mosaics

that have different depths of penetration. These two

systems were compared with parts of the TOBI

(Towed Ocean Bottom Instrument) side scan sonar

data collected on the Irish Margin by the Southampton

Oceanography Centre (S.O.C.) as part of the ENAM

II project (see Stoker and Hitchen, 2003 for ENAM II

objectives and results summary). Sediment was recov-

ered using a 6 m gravity corer. Usually, areas of high

backscattering on the side scan sonar mosaics were

targeted, to examine sand distribution at or near the

sea floor in the canyon area. Such a survey over much

of the length of a deep-water canyon system would

highlight that the mechanisms of sediment transport,

chiefly from sediment gravity flows, have changed

over time as the canyon has progressively shut down.

This would thus allow testing of the simple assump-

tions that are routinely made about canyon backfilling

and shutdown in response to relative sea level rise.

2. Geological setting

The Rockall Trough is situated to the west of

the Hebridean and Irish Shelves on the NE Atlantic

European Margin; the deeper parts of the basin are

mostly within Irish territorial waters (Fig. 1). It is a

major NE–SW trending sedimentary basin that

extends as far to the north as the Wyville–Thomson

Ridge, NW of the Faeroes, where it is 1000 m

deep. It increases in depth to the SW to 4000 m

as it opens out into the Porcupine Trough. The

Rockall Trough is between 200 and 250 km wide,

and is flanked to the east by the Hebridean and Irish

shelves, and to the west by the Rockall Plateau and

adjacent smaller banks to the north (Stoker and

Hitchen, 2003). The deeper part of the Rockall Trough,

below the 2000 m isobath, extends across approxi-

mately half of the basin, from the late Cretaceous–

early Palaeogene volcanic edifices of the Anton Dohrn

and Hebrides Terrace Seamounts, to the SW. The

complex configuration of plateaus, troughs and sea

mounts on this highly topographically irregular

ocean margin is a reflection of Mesozoic–Cenozoic

rifting that led to the opening of the NE Atlantic Ocean

in the early Eocene. This phase has been overprinted

by later, post-Eocene subsidence, Oligo-Miocene com-

pression associated with the Alpine orogeny and late

Neogene uplift (Dore et al., 1999; Naylor et al., 1999).

The study area is located on the northeastern mar-

gin of the Rockall Trough on the Irish Margin to the

west of the Malin Shelf, offshore Eire. This region is

on the dividing line of a major change in the mor-

phology of the deep-water slopes on that basin margin

flank. To the north of the study area the Hebridean and

east Rockall Bank shelves have seen considerable

uplift since the early Pliocene, with the shelf margins

having prograded by up to 50 km (Stoker and

Fig. 1. Location of the northeastern Rockall Trough margin and study area. Bathymmetric data source: GEBCO. SSF: Sula Sgeir Fan; BF: Barra

Fan; DF: Donegal Fan; RKB: Rockall Bank; N.I.: Northern Ireland; Smt: Seamount.


Hitchen, 2003). This uplift was related to the regional

Miocene uplift that also strongly affected the northern

North Sea. The Sula Sgeir Fan off the NW Hebridean

shelf, the Barra Fan southwest of the outer Hebrides

(north of the Hebrides Terrace seamount) and the

Donegal Fan (immediately to the north of the study

area) are the most prominent of these progradational

wedges, and record extensive shelf glaciation since

the mid-Pleistocene (Stoker and Hitchen, 2003). To

the south of the study area towards the western flanks

of the Porcupine Bank, the slope has a destructive

profile and is characterized by mass-wasting that has

persisted since the late Palaeogene (Stoker et al.,

2001). The complex topography that resulted from

this structural configuration of elements implies that

the entire margin has had a profound influence on

palaeoceanographic circulation and deep-water sedi-

mentation. The eastern margin of the Rockall Trough

in particular has been constructed by a combination of

along slope and downslope processes, and is referred

to as a sediment loaded (in contrast to a sediment-

starved) slope, with turbidite fans (e.g. Sula Sgeir Fan,

Barra Fan, Donegal Fan; Fig. 1) overlying and inter-

digitating with older and contemporary along slope

sediment drift deposits (Stoker and Hitchen, 2003).

Along-slope transport of sand is known from the

upper slope in this area (Kenyon, 1986).

The general morphology of the southern Rockall

Trough on the Irish margin is known from previous

surveys. The slope is characterized by a series of

canyons which cut back almost to shelf break, and

another series of associated canyons which appear to

initiate part or halfway down the continental slope.

These canyons have areas of high backscatter associ-

ated with their floors. Apart from the rare but often

dramatic canyon and channel features, the slope is

characterized by a series of features indicative of

mass wasting. These include slumps, slides, slump

scars, failure terraces (oriented parallel to slope), and

other more enigmatic features that could be inter-


preted as sedimentary bodies beyond the canyon

mouths, or bodies scattered across the slope, of un-

known origin or composition.

The Donegal Bay Canyon (first named here) is the

focus of examination within the study area, and is the

most prominent submarine canyon surveyed in detail

to date in the region. The head region is located at

approximately N55845VW10840V, below the shelf

break to the west of Donegal Bay in the Republic of

Ireland. Its existence has been known for some time,

and it is shown as a marked wide embayment into the

shelf edge profile in many atlases (e.g. Times Books,

1999).

3. OKEAN side scan sonar and 3.5 kHz profile

data

The study area was surveyed using OKEAN, a

Russian medium-range sidescan sonar tool, a hull-

mounted 3.5 kHz profiling system, and simultaneous-

ly with multichannel seismic. Fig. 2 shows the profile

tracks over the study area from the upper to lower

slope. Fig. 3 shows the OKEAN side scan sonar

mosaic, with high backscattering areas shown in

light shades, and lower backscattering areas shown

in darker shades. Cores locations are also highlighted

and will be discussed below in the ground-truthing

section. Fig. 4 shows an interpretation of the

OKEAN mosaic, which clearly shows the Donegal

Bay submarine canyon. The head area is impressive,

with a large dcauliflowerT shaped amphitheatre up to

20 km in diameter, fed into by a network of incised

tributary gullies just below the shelf break. The

OKEAN shows this area to be broadly highly back-

scattering in character, interpreted to correspond to

mud-poor and sand-rich sediment at or near the

canyon floor. The middle and lower parts of the

OKEAN mosaic are medium backscattering in char-

acter, and this is interpreted to reflect depositional

topography. This part of the slope shows one main

trunk canyon running downslope from the complex

head region.

The 3.5 kHz hull-mounted profiling survey was

carried out simultaneously with the seismic and

OKEAN data acquisition, at a speed of 6 knots.

The data were of superb quality, and were used

initially for the siting of gravity core locations. Sub-

sequently the profiles were used more systematically

to describe the shallow subsurface and the sea floor

topography. Fig. 5 shows the 3.5 kHz lines from the

entire survey area (Fig. 2) from just below the shelf

break to the lower slope area. This figure is thus a 3-

dimensional figure down the complete Donegal Bay

Canyon. A series of 15 acoustic facies were ob-

served and a scheme of their characteristics is pre-

sented (Fig. 6). In the section below, the main

features are described and interpreted. A generalized

map of the acoustic facies distribution is shown in

Fig. 7.

The survey area has three different characteristic

geographic regions of sea floor with associated acous-

tic facies, specific features and sediment recovery

from core. These three areas are (i) erosional area;

(ii) erosional and depositional area; and (iii) deposi-

tional area (Fig. 4).

3.1. Erosional area (lines PSAT 160, 159, 163, 162)

3.1.1. Erosional canyon

Figs. 4 and 5 show that the upper headward parts

of the Donegal Canyon are characterized by evidence

for erosion-dominated processes. Fig. 7 shows that

from 1000 to 2000 m water depths, the acoustic facies

are dominated by 3A, 4A, 5 and 6, which are facies

indicative as hard bottom or outcrop, separated by

areas of parallel acoustic bedding, which increase in

width down system. From 1000 to 1500 m the sea

floor is dominated by closely spaced, V-shaped gullies

separated by ridges. The gullies become tributary to

one another within a broader deep-water valley at

1500 m. The valley width expands rapidly to almost

20 km with a maximum depth of 625 m at this zone of

confluence of the V-shaped gullies. At 1800 m water

depth the sea floor is dominated by one trunk valley

with a width of just over 5 km, and with associated

levees, overbank area, terraces and valley thalwegs.

The valley has the typical characteristics of a

dsteerheadT channel profile. It is ca. 450 m deep

with walls of poor acoustic response.

3.1.2. Intervening ridges

Between the gullies and trunk valley area are in-

tervening areas that are characterized by a variety of

acoustic facies. In the erosional head region the ridges

have internal reflections (e.g. Fig. 5, PSAT 160, 159).

22:34

23:00

23:30

0:00

0:29

1:00

1:30

2:00

2:30

2:583:02

22:00

3:30

4:00

4:234:27

5:00

5:30

6:00

6:18

6:45

7:00

7:30

8:00

8:30

8:40

9:00

9:30

9:36

PSAT-164

23:560:00

0:30

1:00

1:30

PSAT-165

PSAT-161

1:37

2:00

2:30

3:00

3:37

PSA

T-166

4:10

4:30

5:00

5:30

6:00

6:17

PSAT-1

67

PSAT-1

62

PSAT-1

63

PSAT-1

59

PSAT-1

60

PSAT-1

58

6:30

7:00

7:27

PSAT-168

7:33

8:00

8:30

9:00

9:309:32

PSAT-1

69

10:05

10:30

11:00

11:30

12:00

12:14

PSAT-1

70

12:2012:30

13:00

13:29PSAT-171

13:35

14:00

14:30

15:00

15:30

15:44

PSAT-1

72

16:12

16:30

17:00

17:30

18:00

18:30

PSA

T-173

18:35

19:00

19:20PSAT-174

19:26

19:30

20:00

20:30

21:00

21:30

21:44

PSAT-1

75

11:30

12:00

12:30

13:00

13:30

14:00

14:30

15:00

15:30

16:00

16:30

17:00

17:30

18:00

18:30

ORAT-39

ORAT-39

11°30’W 11°20’W 11°0’W 11°00’W 10°50’W 10°40’W 10°30’W

11°30’W 11°20’W 11°0’W 11°00’W 10°50’W 10°40’W 10°30’W

54°40’N

54°50’N

55°00’N

55°10’N

AT282G

AT283G

AT284G

AT285G

AT286G

AT287G

AT288G

AT289G

AT290G

AT291G

AT292G

AT293G AT294G

AT295GAT296G

OKEAN 10 kHZ coverage

OREtech 100 kHZ coverage

AT286G Sampling sites

0 10 20 30

Km 1000

2000

Fig. 2. Map of the Irish Margin with profile (seismic, OKEAN side scan sonar and 3.5 kHz sub-bottom profiler), gravity core and O.R.E.Tech

(deep-towed side scan sonar) locations. Bathymetry from S.O.C. unpublished data.


Those ridges bordering the main Donegal Canyon

show typically strong acoustic response with moder-

ate to deep penetration (facies 1D). The ridges thus

have a constructional appearance that contrasts with

the erosive nature of the gullies and canyons. The

ridges are interpreted to be a combination of erosional

remnants of slope sediments into which the gullies

and canyons have incised, and local evidence of levee

build-up.

3.1.3. Transparent lenses

There is evidence for transparent lenses on the

acoustic profiles in the erosional head area, though

these are restricted to PSAT 159, within the main valley

profile, where they have a hummocky appearance and

laterally restricted occurrence (facies 6). The lenses

have a transparent seismic character and are associated

with local slope changes. They aremuchmore common

in lower reaches of the system. They are interpreted as

11°30’W 11°20’W 11°10’W 11°00’W 10°50’W 10°40’W

10 °40’W10°50’W11°10’W 11°00’W11°20’W11°30’W

10°30’W

10°30’W

54°40’N

54°50’N 54 °50’N

55°00’N

55°10’N 55°10’N

54°40’N

55°00’N

0 10 20 30

Km

AT282GAT282G

AT283GAT283G

AT284GAT284GAT285GAT285G

AT286GAT286G

AT287GAT287G

AT288GAT288G

AT289GAT289G

AT290GAT290G

AT291GAT291G

AT292GAT292G


AT295GAT295G

AT296GAT296G

1000

2000

Fig. 13C

Fig. 10C

Fig. 10A

Fig. 3. OKEAN mosaic from the Irish Margin, showing the location of lines and cores. Light shades are high backscatter, dark shades are low

backscatter. Gravity core locations are shown in white. Location of O.R.E.Tech profile 39 shown in white. Interpretation shown in Fig. 4.


either slumps from the valley walls or as muddy debris

flow deposits filling an older, wider valley, that have

subsequently been eroded near the thalwegs.

3.1.4. Mounded features

A fourth feature observed on the erosional area of

the slope was that of transparent mounds on the

terrace of the Donegal Canyon. The best of these is

seen on the northern side of PSAT-162 at a water

depth of about 2100 m (Fig. 6, facies 6). This

mound is internally transparent on the acoustic profile.

Its origin is unknown.

3.2. Mixed erosional-depositional area (lines PSAT-

167, 166, 170)

The second geographic area is one with evidence

for combined erosion and deposition (Figs. 4 and 5).

284

283

285

292 294

296295293

291

288289290

282

ORAT-39

286

287

sand prone mud pronedepositional

Erosional

Mixederosional

depositional

Depositional

5 kmS H E L F B R E A K

'Cau

liflo

wer

' - s

hape

d he

ad r

egio

n

Fig. 4. Interpretation of OKEAN mosaic shown in Fig. 3.


The main Donegal Canyon is one distinct trunk feeder

channel between water depths of 2200 and 2500 m,

with smaller channels feeding in from the south and

north. There is evidence for hard sea floor bottom

locally, particularly on the floor of the canyon and the

smaller channel. The canyon loses topographic ex-

pression rapidly, down to less than 100 m deep.

3.2.1. Canyon

The Donegal Canyon has reduced in depth, has a

flat bottom, with rough sea floor topography including

remnants of the ridges seen in the erosional part of the

upper slope. The canyon is less than 150 m deep and 4

km wide. Two of the canyons which are seen at the

edge of the survey area (PSAT-166, Fig. 5) are filled

900

1000

1100

1200

1600

1700

2100

2200

2600

2300

2700

2400

2800

2500

1800

1900

2000

2000

1300

1400

1500

1600

1700

2100

2200

2600

2300

2700

2400

2800

2500

1800

1900

2000

2000

1400

1500

2100

2200

2600

2300

2700

2400

2800

2500

2000

2000

2100

2200

2600

2300

2700

2400

2800

2500

2000

2600

2300

2700

2400

2800

2500

2600

2700

2400

2800

2500

2600

2700

28002600

2700

2800

2700

2800

2700

2800

2700

2800

Pr. 160

Pr. 159

Pr. 163

Pr. 162

Pr. 167

Pr. 166

Pr. 170

Pr. 169

Pr. 173

Pr. 172

Pr. 175

N

Erosional

area

Erosional and

depositional area

Depositional area

Canyondirection

Transparent lensmudflow deposit

Main canyondirection

KEY

Dep

th, m

ORAT-39

Fig. 5. Three-dimensional view of the northeastern Rockall Trough slope from shelf to rise (times in milliseconds TWTT). All panels are

interpreted from 3.5 kHz profiles (see Fig. 2).


1A

1B

1C

1D

2A

2B

Facies Schematic Description Type example

Deep penetration; manysimilar reflectors;

arithmetic deterioration ofreflectors

Deep penetration; strongreflectors in middle;

irregular reflectordeterioration

Deep penetration; one hardreflector; several feint,

continuous reflectors; deeperdispersal/diffusion pattern

Deep penetration; arithmeticdecrease in reflector strength;discontinuous lower reflectors

Transparent lens withindistinct base

Transparent lensunderlain by strong

reflectors

Transparent lens withdistinct base

Shallow penetration 1-2hard reflectors

Shallow penetration withrapid decline in reflector

strength

2C

3B

3A

PSAT 168

PSAT 175

PSAT 172

PSAT 172

PSAT 166

PSAT 166

PSAT 164

PSAT 159

PSAT172

Damuth and Hayes, 1977 DescriptionFacies

IA

IB

IIIC

IIIB

IIIA

IIB

IIA

Continuous, sharp bottom echoeswith no sub-bottom reflectors

Continuous, sharp bottom echoeswith continuous, parallel sub-bottomreflectors

Semi-prolonged bottom echoes withintermittent zones of semi-prolonged,discontinuous, parallel sub-bottomreflectors

Very prolonged bottom echoes withno sub-bottom reflectors

Large, irregular overlapping orsingle hyperbolae with widelyvarying vertex elevations abovethe sea floor

Regular single or slightly overlappinghyperbolae with conformablesub-bottom reflectors

Regular overlapping hyperbolae withvarying vertex elevations above thesea floor

4A Rough sea floor; hardand diffuse returns;

diffractions

4B Rough sea floor, lessdark returns; diffuse PSAT 162

PSAT 166

Schematic Description Type example Damuth and Hayes, 1977 DescriptionFacies

6 PSAT 162

5 PSAT 163

7 PSAT 160

8 PSAT 166

Transparent sea floor;transparent mounds

Small blocks above mainSea floor reflector

Steeper slopes; variousInternal reflectors, some

oblique

Hard returns in channels;multiple diffraction

IIIF

IIIE

IIID

Regular, intense, overlappinghyperbolae with verticesapproximately tangent tosea floor

Zones of irregular, intense,overlapping hyperbolae withvertices tangent to the sea floorwhich are interrupted by zones ofdistinct echoes with parallel sub-bottom reflectors

Broad, single, irregular hyperbolaewith disconformable, migratingsub-bottom reflectors

(A)

(B)

Fig. 6. Acoustic facies scheme for the Irish Margin in the vicinity of the Donegal Bay Canyon (left), with Damuth and Hayes (1977) for

comparison (right).


11°30’W 11°20’W 11°10’W 11°00’W 10°50’W 10°40’W 10°30’W

54°40’N

54°50’N

55°00’N

55°10’N

X

X

XXX

XX

XXXXXX

X

X

XX

XXXXXX

XX

X

XXXXXX XX

XXXX

XXXXXXXXX

XX

LEGEND:

PARALLEL ACOUSTIC BEDDING (Facies 1, 3B)

TRANSPARENT LENSES (Facie 2)

HARD BOTTOM (Facies 3A, 4A, 5)

OUTCROP (Facies 3A, 6)

TRANSPARENT HUMMOCKY (Facies 6, 4B)

XXXXX

1000

2000

500

2500

0 10 20 30

Km

Fig. 7. Map showing the generalized distribution of 3.5 kHz acoustic facies.


with transparent acoustic facies (facies 2B; locally

facies 2A and facies 2C). The main Donegal Canyon

has higher acoustic penetration (facies 1B).


The intervening low-relief ridges seen on PSAT-

166 in the middle of the erosional–depositional area

have small transparent lenses, in addition to those

recognized at the bases of the smaller feeder canyons.

These lenses are found predominantly on the slopes of

the ridges. Some of the lenses seen on the canyon

floors can be traced laterally, updip, onto flatter or

steeper areas of slope. These lenses are interpreted

here as muddy debris flow deposits.

3.3. Lower Slope area (lines PSAT 169, 173, 172,

175)

This geographic area includes the lower reaches of

the Donegal Bay Canyon. Depths range from 2595 to


2665 m, which is a relatively flat area compared with

the other two areas. A generally strong reflector char-

acter with rough topography is observed, with either

parallel acoustic bedding (facies 1, 3B) or transparent

lenses (facies 2).


The lower reaches of the Donegal Bay Canyon are

characterized by laterally extensive transparent

acoustic facies lenses, which are locally sheet-like

in geometry, and are concentrated in topographic

lows. These lenses are seen on the lower profiles in

Fig. 5. Typically the lenses pass laterally or pinchout

onto topographic highs with facies 3A acoustic char-

acter. Lenses are present at depth ranges between

2650 and 2740 m, and are of variable thickness.

The thickest lens (b7 m) is partially on a slope to

the east where it onlaps or passes into diffuse and

then sharp reflectors, though several of the lenses are

seen to abruptly terminate against areas with very

sharp, sometimes even vertical terminations. One of

the wider lenses is 2 km wide. This lens occupies the

much shallower (75 m deep) down dip reach of the

Donegal Bay Canyon.

The transparent lenses are interpreted as canyon

mouth depositional areas occupied by slump or deb-

rite deposits and perhaps coarser material, bordered

on the slopes on either side by hemipelagic material.

The debris flow deposits appear to infill an eroded

canyon topography. There is evidence for erosion

here with a terraced erosive cut that is filled by

debrite material, which is probably muddy. The west-

ern margin of the valley appears to be associated

with faults.

4. O.R.A.Tech side scan sonar data

The O.R.A.Tech side scan sonar is a high-resolu-

tion system that was operated at 100 kHz providing a

swath of 1000 m (500 m each side). The vehicle was

towed at about 50 m above the seabed. The

O.R.A.Tech system also includes an acoustic sub-

bottom profile (SBP) that operates at 7 kHz in order

to give good resolution.

One O.R.A.Tech profile was designed on the

basis of backscatter characteristics of a previous

TOBI (towed ocean bottom instrument) survey in

the area (S.O.C., unpublished data), and information

gathered from the OKEAN mosaic collected during

this leg (Fig. 3). High backscattering features on the

TOBI were thought to be sand or gravel features,

within a broader canyon morphology. An exercise in

ground-truthing and comparison of side scan sonar

tools with different depths of penetration and reso-

lution was undertaken to resolve the enigmatic fea-

tures, and this is discussed in more detail in the

next section.

4.1. ORAT 39 description

The O.R.A.Tech deep-towed 100 kHz side scan

sonar provided detailed (high resolution) imaging of

the area surveyed. Line 39 (Fig. 2) is positioned in the

lower reaches of the down slope-running canyon sys-

tem on the continental rise of the Irish margin. The

line runs from 55804VN–10857VW to 54852VN–11812VW in a NE–SW direction. It is about 24 km

long and located halfway between the 3.5 kHz profile

and OKEAN/seismic lines PSAT 169 and PSAT 170

(Fig. 3). The positioning of ORAT 39 was based upon

the previous TOBI 30 kHz side scan mosaic in which

some of the acoustic facies were difficult to interpret

(Fig. 8).

On the basis of the O.R.A.Tech profile it is noted

that topography is flat, denoting a broadly aggrada-

tional relief. The depths range from 2580 to 2680 m.

The 2580 m depth corresponds to a topographic

elevation, which is a wide, gently sloped ridge run-

ning SE–NW as seen on the bathymetric chart (Fig.

2), on the northeastern section of the line. The 2680

m depth is correlated to the thalweg of a V-shaped

channel partly infilled with sediments of low acous-

tic backscatter that give the channel its current U-

shaped profile. It is located on the mid-section of the

line, together with five other channels of higher

backscatter and reflectivity on the profile. Four of

them are U-shaped, with very low angle walls, and

one is V-shaped with an asymmetric profile. This

system of channels is positioned on a topographic

depression correlated to the main canyon in the

upper reaches of the area surveyed by OKEAN

and 3.5 kHz profile.

Towards the southwestern end the depths gently

decrease to 2650 m, corresponding to a gently sloped

wide ridge, running SE–NW from the upper reaches

5

7

11

3

1a9

1a

3

11

7

6

5

6

NE11:34

SW15:20

Start

SW

NE15:20

18:45

End

2600m

2550m

2600m

2550m

1a

4a

2

1b

3

4a

9

8

10

11

3

9

4b

10

9

700 m

700 m

Fig. 8. Mosaic and facies outline of O.R.E.Tech line ORAT-39—Rockall Trough. Acoustic facies are shown in Fig. 9, and described in the text.

See Figs. 2–4 for location.



of the area. The declivity between the deepest chan-

nels and the NE ridge is 0.18 or 1:150 (the bathym-

etry increases 1m for every 150 m). This is twice

that of the declivity between the same channel thal-

weg and the southwestern ridge, which is 0.178 or

1:300.

(a)

(b)

Parallellineations of mediumbackscatter in a homogenous areaof low backscatter.

as with subtle lineations(a)

(a)

(b)

Slightly rough texture; smallstipples of medium backscatter;low scatter background.

as with parallellineations(a)

low

back

scat

ter

high

back

scat

ter

2

1a-b

3

4a-b

6

7

8

9

10

11

DescriptionExampleFacies

Elongate patches of mediumbackscatter on a smoothbackground of low backscatter

Smooth, homogenous areaof lowbackscatter;little acoustic variation(occasional stippling)

Smooth; homogenous; featureless;medium backscatter

Slightly rough texture; stippledhigher backscatter;mediumbackscatter background

Rough, granulartexture; highbackscatter

Distinct features; varying width;very high backscatter;roughtextures

Irregularzone; high backscatter;occasional parallellineations; somegrooves and ridges

Discontinuous linear features; verynarrow;fine-medium texture

Scattered lineations; highbackscatter;coarse granulartexture; dispersive

5

med

ium

back

scat

ter

line

arfe

atur

es

Fig. 9. Facies scheme for O.R.E.Tech 100 kHz line—ORAT-39

(Fig. 8).

4.2. ORAT 39 acoustic facies description

Eleven acoustic facies have been recognized on

line ORAT 39 (Fig. 9). They have been classified

into four groups according to their backscatter re-

sponse and the presence of linear sea floor shapes.

The low backscatter group comprises facies 1a, 1b, 2,

3, 4a and 4b. The medium backscatter group is made

up of acoustic facies 5 and 6. Facies 7, 8 and 9 are

within the high backscatter group. The final group

comprises acoustic facies 10 and 11 that have different

patterns of linear features. All of the facies are dis-

tinctive and readily recognized. This side scan sonar

line is particularly interesting because of the large

number of contrasting backscatter patterns, whose

origins are rather enigmatic. Ground-truthing of

these acoustic facies is discussed in the next section.

The polarity on the images is positive which

implies high backscatter response is in black–dark

grey colour, and low backscatter is in light grey–

white. Fig. 8 shows the acoustic facies/backscatter

patterns along line ORAT 39. The facies can be

described as wide, linear features of approximately

180 m width. The wide, linear facies may occur

either orientated SE–NW, normal to the line (facies

1a, 1b, 2, 3, 4a, 4b, 5, 6 and 7) or oriented obliquely

(608) to the line (facies 8, 10 and 11). Only facies 9

has two different orientations. It either occurs as a

narrow, linear feature (30–200 m wide on average)

oriented both SE–NW (normal to the line) and E–W

(oblique to the line at 308 and 608) or as an irregular,

U or horseshoe-shaped (in plan view) patch, opening

to the NW.

5. Ground-truthing

A series of gravity cores were collected on the

basis of backscatter acoustic facies from the

O.R.A.Tech profile ORAT 39 and its 7.5 kHz profile,

and the 3.5 kHz profile sections (Figs. 10–13). In this

section we discuss the gravity cores from three differ-

ent geological/physiographic parts of the Irish Margin

deep-water slope, and see how the sedimentary layer-

ing may explain the backscattering on the OKEAN

and O.R.A.Tech sonographs. These three areas are: (i)

Canyon Head region; (ii) Lower Slope region and (iii)

Upper Rise region. In all regions, a pattern-coding of

Fig. 10. (A) Sedimentary core logs (depth in cm) AT-282 to AT-285G with positions on (B) 3.5 kHz hull mounted profiler line PSAT 162 and (C) on OKEAN side scan sonar mosaic.

AT-282 hit buried turbidite sand in the overbank area; AT-283 hit sand on the sea floor on a mound on the terrace; AT-284 and AT-285 both hit dstickyT debris flow that had slumped in

from the side. AT-286 was outside the line of section. See Fig. 2 for location.

B.T.Cronin

etal./Sedimentary

Geology179(2005)71–97

86

Fig. 11. 7.5 kHz O.R.A.Tech profiler line ORAT-39 (see Fig. 3) with core logs (depth in cm) AT-292 to AT-295G. AT-292 and AT-293G

penetrated muddy debrites. AT-294G penetrated into shallow turbidite sands in a narrow channel seen on the profiler. AT-295G penetrated a

wide, empty channel. Core AT-296 was not on the profile.


the core numbers at their locations (as shown on the

sonograph details) is used. A stippled pattern is used

for those cores where sand (of both turbidity current

and bottom-reworking/contour current origin) was re-

covered, a grey pattern for reworked mud (both

muddy debris flow and slump origin), a pebble pattern

for prominent layers of clast supported gravel of

enigmatic or compound origin, and a white pattern

for hemipelagic sediment. The same patterns are used

in the core logs (Figs. 10–13).

Of the two side scan systems used on this cruise,

the OKEAN system with a frequency of 10 kHz has a

maximum penetration of 10 m, but lower resolution

than deep-towed systems. The O.R.A.Tech has a dual

frequency but in this ground-truthing exercise we

operated the 100 kHz band, which results in penetra-

tion of up to approximately 0.3 or 0.4 m, and has the

highest resolution of the two. TOBI (not used on this

cruise) may resolve features that are too deeply buried

for O.R.A.Tech to identify.

5.1. Canyon Head region

Six gravity cores were taken in the Canyon Head

region: four in a transect across the upper reaches of

the main deep-water canyon (AT-282G to AT-285G),

one from the floor of the same canyon approximately

8 km nearer the shelf break up the canyon axis (AT-

286G) and one core (AT-287) was taken from a

mound off the canyon flank in the extreme SW of

the surveyed area, approximately 17 km up dip from

the canyon transect (Figs. 2 and 10C).

AT292GAT292G

AT293GAT293G


AT296GAT296G

2 km

- mud flow - clast-supported layer - sand - hemipelagic

AT-292GAT-293G AT-294G AT-295G

OKEAN (10kHz) O.R.E.tech(100kHz)

(A)

(B)

N

50m

700 m

Fig. 12. (A) Comparison of two different side scan sonar tools operating at different frequencies. At 10 kHz, OKEAN may penetrate several to

10s of meters below the sea floor—thus buried features are frequently imaged that drop cores may not reach. At 100 kHz, O.R.A.Tech will

penetrate only up to 0.2/0.3 m, and thus only very shallow or sea floor features are imaged. Backscattering penetration is tested by targeting

features with drop cores. See text for details. (B) O.R.A.Tech profile showing core locations. Compare with Fig. 11.


5.1.1. Canyon transect

Fig. 10C shows the core logs, the OKEAN data in

the canyon transect and PSAT 162 3.5 kHz profile

data in the immediate vicinity of the transect with the

positions of the cores.

Core AT-282G was targeted in an area of multiple

layering on the 3.5 kHz profile data which is outside

the canyon, and is thought to be the overbank area to

that canyon. The backscattering on the OKEAN is

medium to low. A linear feature at the core location

running NW–SE and then doglegging towards the

canyon and is probably a buried channel or canyon

(Fig. 10C). The core is not thought to penetrate this

object. The core recovered bedded hemipelagic sedi-

ment and three sands between 2.4 and 3.5 m that

contain siliciclastic material with foraminifera and

local glacial dropstones.

Core AT-283G was targeted on a topographic high

between the overbank and canyon axis areas, and is

thought to be the highest part of the canyon levee (or

erosional remnant, if not aggradational), with a trans-

parent mound structure clearly visible on the 3.5 kHz

data. The OKEAN sonograph shows medium–high

backscatter, although this is somewhat obscured by

the ship’s-track and by a nearby acquisition anomaly

on the sonograph. The core location is near the in-

ferred buried channel at core station AT-282G. The

core recovered bedded sands and foraminiferal ooze,

though it is dominated throughout by sand. Penetra-

tion was poor (0.8 m), and this was thought to be due

to the core barrel hitting deeper, impenetrable sand

below this level.

Core AT-284G was targeted on the deepest part of

the canyon, i.e. the inferred canyon thalweg, where

the 3.5 kHz profile shows a hard, shallow-penetra-

tion acoustic response with some diffractions. On the

OKEAN, the backscatter is medium and the sedi-

mentary body is clearly part of the bcauliflowerQ-shaped area of the Canyon Head region. The core

recovered hemipelagic sediment and hit the top of a

muddy slump or debris flow interval, though core

penetration was poor (0.6 m). This area corresponds

to the location of two large transparent lenses within

the canyon.

Core AT-285G was targeted on a mounded fea-

ture on the canyon floor with highly irregular sea

floor topography, strong acoustic returns on the side

scan sonar, and diffractions on the 3.5 kHz profile.

Fig. 13. (A) Sedimentary core logs (depths in cm) AT-288 to AT-290G with (B) 3.5 kHz hull mounted profiler line PSAT 175 with core positions, and (C) core positions shown on the

OKEAN side scan sonar mosaic. AT-290 penetrated a debris flow with pebbles, corresponding to the transparent lens on the 3.5 kHz profile; AT-289 hit the feather edge of the debris

flow floored by turbidite sand, and AT-288 pelagic marls.

B.T.Cronin

etal./Sedimentary

Geology179(2005)71–97

89


The feature was interpreted as a slump mass on the

canyon floor. On the OKEAN the backscattering is

medium, a little higher than at station AT-284G, but

clearly in the axis of the main canyon at the down-

dip end of the bcauliflowerQ structure. The core

recovered a hemipelagic sequence with intercala-

tions of slumped material. The slumped intervals

are dewatered and extremely sticky in texture, mak-

ing it difficult to extract the sediment from the

core-liner.

5.1.2. Core AT-286G: Canyon Head core (Fig. 10)

Core AT-286G was taken approximately 7 km up-

canyon towards the headwall of the cauliflower pat-

tern imaged on the OKEAN mosaic (Figs. 3 and 10).

The core station was not targeted on the profile, but it

does coincide with the sonographs. On the OKEAN,

the location has low backscatter within the central part

of the cauliflower structure. A deep canyon with high

backscatter margins is seen clearly within the cauli-

flower. The core was targeted at the axial part of the

canyon floor, and recovered intercalated thick sandy

layers (the thickest recovered on the margin) and

hemipelagic sediments.

5.1.3. Core AT-287G: marginal mound feature

Core AT-287G was targeted using the 3.5 kHz

profile on a mounded structure on the flank of the

main canyon further up canyon from AT-286G. The

structure, despite its presence at an unusual depth

(~1300 m), was targeted as a possible carbonate

mound. Such carbonate mounds are commonly

found on the Porcupine Trough margin but at shal-

lower depths (~500–1200 m: Kenyon et al., 2003),

though Lophelia pertusa, the dominant coral species

that is found on these mounds, is known in water

depths up to 2000 m in the Atlantic (Le Danois,

1948). The structure is one of several features of the

same size seen on PSAT-161, at water depths between

1240 and 1280 m. On OKEAN, the core station is in a

localized area of high backscatter situated on the flank

or just outside a major SSW–NNE trending canyon

that feeds into the SW part of the bcauliflowerQ struc-ture. There was no recovery in the gravity corer, but

there were sand and clasts (interpreted as dropstones)

in the core catcher, suggesting that the features are

more likely to be sediment waves than carbonate

mounds.

5.1.4. General interpretation of the Canyon Head

region

The bcauliflowerQ-shaped head area seen so spec-

tacularly on the OKEAN mosaic is not seen on the

shallower penetration TOBI mosaic (S.O.C. unpub-

lished data). On OKEAN, stacked linear canyon and

channel features and tributaries with low backscatter

are interpreted to record major phases of erosion and

sand deposition in the head area that do not occur at

present. On 3.5 kHz profiles the dominant sedimen-

tary bodies overprinting this older phase are slumps

and muddy debris flows. Most of the smaller-scale

mudflows and slumps are directed towards the canyon

and channel axes, away from the canyon walls. This

indicates that the system does not transport sand at

present. One slump body seen on TOBI and cored

twice is thought to have blocked the main canyon

axis, and subsequent sand transportation may have

been ponded behind it. In targeting sand with the

gravity corer, it was found that coring the topographic

highs near canyons, coring areas of low backscatter

near to the head of the system, and collecting samples

from the deepest part of a V-shaped thalweg, are the

most reliable ways. Other potentially sand-rich areas

were too deep below the sea floor (seen on OKEAN)

to be sampled by the 6 m gravity corer, largely due to

the thickness of the slumped masses. The overbank

area in particular (e.g. AT-282G) shows that relatively

recent large sandy turbidity currents moved through

the canyon, and spilled into this region. This is par-

ticularly impressive when one considers the width (17

km) and depth (~800 m) of the canyon. The timing of

these flows is thought to be towards the end of the last

glacial period, confirmed by their association with

dropstones.

5.2. Lower Slope region

Five cores were collected in the mixed erosional–

depositional, lower slope area of the Irish Margin

(Figs. 3 and 11). Cores AT-292G to AT-295G were

all selected on the basis of their backscatter character-

istics on O.R.A.Tech, in combination with informa-

tion about seafloor topography and acoustic response

from the 7.5 kHz profile (Fig. 12). Core AT-296G was

selected based on backscatter patterns. The cores were

taken in a strike (along-slope) transect (Fig. 3). Cores

AT-292G and AT-293G were targeted on depositional


features on the profile, to test the reason for the

difference in backscatter on OKEAN and O.R.A.Tech.

Cores AT-294G and AT-295G were targeted on ero-

sional features. Core AT-296G was targeted on a very

high backscattering area seen on TOBI, where sea-

floor topography was unknown.

Core AT-292G was taken from an area of medium

backscatter on OKEAN and TOBI, and low backscat-

ter on O.R.A.Tech (Figs. 3 and 8). On the profile the

sub-bottom is characterized by deep penetration, mul-

tiple layering. A transparent layer was observed on the

profile, which can be traced laterally under core sta-

tion AT-293G, pinching out before station AT-292G

(Fig. 11). The core recovered a thick mudflow deposit

capped by Holocene marl.

Core AT-293G was taken from an area of medium–

high backscatter on both OKEAN and O.R.A.Tech

(Fig. 12). On the profile the seafloor reflects strongly

but otherwise is the same as station AT-292G. The

core recovered a mudflow deposit capped by a marl,

but separated by a 0.05 m thick clast-supported gravel

(Fig. 11).

Core AT-294G was taken from an area of medium–

high backscatter on OKEAN, a medium–high back-

scattering, narrow, linear feature on TOBI (with a

NW–SE orientation), and an area of low backscatter

on O.R.A.Tech. On the profile this area is in a small,

V-shaped channel with a thin layer of transparent

acoustic response underlain by stronger returns and

poor penetration (Fig. 11). The core had poor recov-

ery, with a Holocene marl capping sand. The marl is

interpreted to correspond to the transparent acoustic

facies on the 7.5 kHz, and the sand to be the top of the

higher amplitude acoustic facies below the channel

base. The O.R.A.Tech does not show the sand (0.38

m) due to lack of penetration.

Core AT-295G was taken from an area of medium

backscatter on OKEAN and low backscatter on

O.R.A.Tech. On the profile the core is from the mid-

dle of a zone of strong reflectivity within a flat-

bottomed, shallow depression. The core recovered a

thick hemipelagic sequence without sand, draped by

Holocene marl. The depression is interpreted as an

inactive, draped channel (Fig. 11).

Core AT-296G was taken from an area of me-

dium backscatter on OKEAN. The core recovered a

thick hemipelagic interval capped by a Holocene

marl, and separated by a thin clast-supported gravel

(reworked dropstone) layer, identical to that seen in

AT-293G.

5.2.1. General interpretation of the Lower Slope area

The different side scan sonar systems allow 3D

analysis of the upper sedimentary layers, in combina-

tion with 3.5 kHz, 7.5 kHz and seismic profiling

across the same transect (Figs. 11 and 12). The high

backscattering gravel layer seen on TOBI (S.O.C.

unpubl. data) is not as evident on OKEAN, because

it is not thick enough, and not as evident on

O.R.A.Tech, because it is partially buried. The same

can be said for sand-filled features such as the narrow

linear channel seen on TOBI, which is not clear on

OKEAN (Fig. 12), but perhaps shows up as a field of

linear features (acoustic facies 10 on O.R.A.Tech, Fig.

9). The transparent lens on the profile to the SW is

attributed to the (cored) mudflow that is buried by the

gravel layer in the centre of the profile. It is a sheet-

like lens extending across half of the profile, and is

thought to terminate at this locality rather than be

removed by erosion. The complex nature of mudflow

deposition, inferred from the profile and cores, cannot

be resolved from the sonographs, but it is inferred

here that there are many, interfingering mudflows

possibly coming from different local directions. This

is in agreement with conclusions from the Canyon

Head region; the system is now dominated by rework-

ing of intraslope mud. The gravel layers sampled at

AT-293 and AT-296 are thought to be of identical age.

Therefore the central part of the profile is interpreted

as an erosive valley (low relief area of erosion) that

truncated the gravel, and possibly the transparent lens

underneath.

5.3. Upper Rise region

Four cores were taken in the Upper Rise area.

Cores AT-288G to AT-290G were targeted on a

range of acoustic facies on the 3.5 kHz profile and

on OKEAN line PSAT-175. This section is a slope-

parallel transect in the depositional area of the Irish

Margin (Fig. 13).

Core AT-288G was targeted on the edge of a very

thin transparent lens on a topographic high otherwise

characterized by medium penetration, multiple layer-

ing (which deteriorates rapidly) in the sub-bottom.

The area is outside the wide, shallow valley that


corresponds to the area downdip of the main canyon

mouth. The core recovered a hemipelagic sequence

with two coarse-grained, clast-supported layers at

0.85 and 0.9 m. These coarse-grained layers are over-

lain by silty hemipelagic sediments, which may cor-

relate with the sands in AT-289G, and the succession

is capped by Holocene marl. Two thin, silty–sandy

layers were recovered towards the bottom of the core.

Core AT-289G was targeted on the edge of the

shallow valley, on an undulose low-angle slope. The

3.5 kHz profile is characterized by the feather edge of

a thin transparent lens underlain by a strong reflector

and some deeper strong reflectors. The core recovered

a thick sequence, with the Holocene marl underlain by

a thin mudflow, underlain by several graded silty–

sandy layers (max. thickness 0.2 m), interbedded

with hemipelagic sediments, a thin, clast-supported

gravel at 1.3 m, and a thick package of hemipelagics

without gravel. Several thin silty layers are seen be-

tween 2.5 and 2.7 m.

Core AT-290G was targeted on an area of slightly

higher backscatter on the OKEAN, where the trans-

parent lens on the 3.5 kHz profile is thicker, and

nearer the deepest part of the shallow valley than in

core AT-289G. The core recovered a thick sequence,

with the Holocene marl underlain by a 3.7 m thick

mudflow, with various layers of gravel, some clast-

supported, at its lower parts. Due to their isolation in

mudflows or in situ deep-water muds, and distance

from potential source of such rounded gravels, they

are interpreted as dropstones. In situ Holocene sedi-

ments underlie the mudflow.

6. Discussion

Submarine canyons are a major feature of deep

marine slopes, and act as transfer zones for clastic

sediment to deeper water. Fan models usually place

the submarine canyon as the point source for the fan,

and other areas of the slope are typically mud prone

with mass-wasting being the main process of sediment

remobilisation. The Donegal Bay Canyon is an exam-

ple of such a sediment transfer zone from the north-

western European margin. The morphology of the

canyon was examined using a variety of methods,

including high resolution 3.5 kHz seismic profiling,

low and high-resolution sidescan sonar imaging, and

ground truthing of the acoustic facies using gravity

corers. The canyon was examined from its head area

down to the continental rise. The head area is a classic

example of a dcauliflowerT-shaped, head area, previ-

ously described as an damphitheatreT-like head area byother workers (e.g. Kenyon et al., 1978). This head

region is the widest part of the canyon system, and

comprises a confluence area of a series of V-shaped

gullies, each with dimensions in the region of 100–

500 m wide and 150–300 m deep. The confluence

area is one of highly irregular sea floor, with multiple

V-shaped incisions within the larger canyon fairway.

Downdip the canyon passes rapidly into a single U-

shaped fairway that narrows to 2–3 km, with a depth

of 300 m. Downslope, the canyon is joined by several

other tributary canyons of varying dimensions, before

passing into a broad, shallow, scoured area and then

losing topographic expression on the continental rise.

The margins of the canyon are characterized in the

upper and medial reaches by terracing, giving the

canyon a dsteerheadT cross-sectional profile, and by

multiple slump scars, giving a dscallopedT morpholo-

gy to the edges of the fairway. This terraced character

and the prevalence of slump scars along the margins

has been described by other authors from slope can-

yons and channel complexes (e.g. Cronin et al., 1995;

Wheeler et al., 2003). The slump scars are interpreted

to reflect lateral wasting of canyon margins towards

the canyon axes by slumping. Terraces are interpreted

as reflecting larger scale wasting of the margins to-

wards the axes, by slow creep caused by rotational

slumping along discrete gravity faults. The surface

expression of this type of dslowT margin failure to-

wards the axis, seen in the OKEAN mosaic interpre-

tation in Fig. 4, is compelling. The implications of this

are far-reaching, as this type of canyon or slope

channel margin terracing, and even the position of

the interpreted faults from this model, is interpreted

as the product of dinside leveeT development in some

current models (Deptuck et al., 2003). In these mod-

els, the dipping reflectors that we interpret as parts of

the rotational slumps, are interpreted as the aggrada-

tional elements of a confined channel levee complex

within the main canyon or slope channel. Further-

more, the scalloped pattern, interpreted as axis-direct-

ed slump scars, is interpreted as the erosional margins

of underfit sinuous channel elements in the same

models. The dinside channelT model does not work


for the Donegal Bay Canyon, and both end-member

models need to be tested more closely for other slope

channel and canyon complexes.

Submarine canyons are thought to initiate by head-

ward erosion of a submarine slump scar in the upper

slope region. As the feature gets larger due to contin-

ued retreat of the headwall, the canyon is modified by

episodes of mass wasting and the headwall retreats

towards the shelf break, where it becomes inclined to

trap turbidity currents spilling over the break. At this

point the canyon becomes a major sediment transfer

zone and if this process continues, the canyon may

erode back across the shelf towards any fluvial or

longshore drift sediment source. Most sediment is

transported to the end of the canyon system by tur-

bidity currents, and larger flows may even spill over

the walls of the canyons in the upper reaches, though

this is usually more common in the lower reaches,

where the canyon may pass into a leveed valley

system. Canyon bypass is the main phase of turbidite

fan growth, and corresponds to the classic Type I

turbidite system of Mutti (1985).

As relative sea level rises, present models infer that

the lower parts of the canyon will become deposition-

al, and that sand deposition is confined to the lower

slope area, perhaps with the fan backstepping into the

canyon mouth and even into the canyon itself. This

corresponds to a Type II turbidite system (Mutti,

1985). Continued sea level rise will see the sand

trapped on the shelf area, unable to be transported

into the submarine canyon as it is locked up-system.

The submarine canyon and slope area will thus revert

to a mud-starved doverbank wedgeT and the canyon

will be mud-plugged. This corresponds to a Type III

turbidite system (Mutti, 1985). Though this scheme

has been modified to take in subsequent concepts such

as flow efficiency, the main tenets remain the same in

the literature.

Fig. 14 is a model proposed to explain the mechan-

isms of filling observed in the vicinity of the Donegal

Bay Canyon. Three stages of activity within the main

canyon area are recognized during this filling period

(Fig. 14).

Stage 1 corresponds to the stages of sediment

bypass through the canyon, inferred to represent the

end of the last glacial low stand, when sand was

probably transported directly from a shallow Malin

Shelf area, where large volumes of sand were present

on the shelf, and initially bypassed the major Donegal

Bay Canyon into deeper water. The canyon base is

strongly erosive, even in the lower reaches, during this

stage. Occasionally sand was deposited adjacent to the

head area of the main canyon, thus spilling onto the

flat upper slope area in general. This has major impli-

cations for flow volumes as the canyon is 17 km wide

and up to 800 m deep in that area.

Stage 2 corresponds to a phase of mixed erosion and

deposition within the Donegal Bay Canyon. Sand is no

longer found outside the canyon in the upper reaches,

and the largest flows only managed to spill onto the

terraces within the main canyon in that area. Sand was

deposited in the upper and medial reaches of the can-

yon, and flows were smaller in volume and usually

confined to smaller channels within the main canyon.

Stage 3 saw a transition from sand and silt trans-

portation to one of muddy debris flow deposit and

slump dominance in the medial and lower reaches of

the Donegal Bay Canyon. These deposits are clearly

seen to overlie erosive features and coarser sediment

from stages 1 and 2 on the 3.5 kHz records. Slumps

were particularly common in the upper reaches, where

they blocked the main canyon trunk axis. Subsequent

turbidity currents that were transporting sand were

ponded in the canyon in the upper reaches, rather

than depositing sand sheets in the lower reaches of

the canyon. This is also complicated by sand spill over

from the shelfal region and from along slope trans-

portation of sediment as coarse sandy contourites

(Kenyon, 1986).

This paper presents a dataset where mechanisms

for canyon filling on a glaciated margin, which has

progressively drowned since the last glacial low stand,

are investigated. After a sustained phase of coarse

clastic sediment bypass through this major erosive

feature on the Irish Margin, the locus of sand deposi-

tion moved into the upper reaches. A later phase of

mass-wasting plugged the canyon and ponded the

sand in these upper reaches.

There are implications for the understanding of

deep-water slope channel complex filling, particularly

in West Africa where they are known to host major

fields of oil and gas (Mayall and Stewart, 2001).

There are many more potential stratigraphic hydrocar-

bon plays in this new model, particularly during stage

3, than predicted in existing models. Current under-

standing of the architecture of slope canyons is that

1

Stage 1:

All bypass (Mutti type I);Some overspilling of sandin head region

erosion

Overbanksand

Sandy terraces

2

Stage 2:

Occasional Bypass;Decrease in volume of flowsin lower reaches

3

Stage 3:

Plug and pond;Slumps from wallsdominate lower canyon fill and pond in uppercanyon

Continental rise(2700 m)

Shelf edge(1000 m)

5 km

Slope (2500 m)

overbanking

Bypass

Bypass and

Channelling, erosion and deposition (mixed)

Bypass

Slump:ponding of sand

Drape, slump,rare channelised sand

Slump anddebris flow

Fig. 14. Model for canyon filling in response to relative sea level rise.

B.T.Cronin

etal./Sedimentary

Geology179(2005)71–97

94


they have complicated multi-phase fills between ini-

tial excavation, bypass and erosion, and eventual

plugging. In all of these models, the dquiescentTphases of activity are always characterized by sinuous

channel elements, which underfill the larger canyons,

and produce sediment fill packages that are highly

heterogeneous and have distinctly lower net:gross

ratios. Sand is found as either erosional remnants of

point bars or as longitudinal, in-channel bars. Slumps

and debris flow deposits are not included in any of

these models. This model implies that longitudinal

slumping and debris flows from canyon margins are

major processes in canyon filling, in distal as well as

proximal canyon reaches. This will produce dmud

podsT of local derivation within canyon fills, and

will also produce sand bodies caused by ponding

like that described in this paper, which are not includ-

ed in current models for slope channel complexes

(Beaubouef and Friedmann, 2000; Campion et al.,

2000; Cronin et al., in press; Mayall and Stewart,

2001; Mayall et al., in press, Wonham et al., 2000).

Acknowledgements

The TREDMAR TTR Program and R/V Logachev

crew are warmly thanked for their support and assis-

tance during data collection on cruise TTR-10. We

wish to thank Jaco Baas, the editor of this special

issue of Sedimentary Geology for this contribution,

for pursuing our revisions of this paper, and pushing it

to completion. BTC wishes to thank the supporting

Oil Companies from the dMesostratigraphy of Deep-

Water SandstonesT consortium: Amerada Hess, BP-

Amoco, Elf, Enterprise, and Conoco; and laterally,

ENI Agip, during the period that the work was carried

out. We wish to thank the following referees for their

considered and thoughtful reviews of earlier versions

of this paper: David Piper, Bill McCaffrey and Jamie

Pringle.

References

Akhmetzhanov, A., Kenyon, N.H., Ivanov, M., Cronin, B.T., 2003.

The continental rise west of porcupine sea bight, Northeast

Atlantic. In: Mienert, J., Weaver, P. (Eds.), European Margin

Sediment Dynamics: Side-Scan Sonar and Seismic Images.

Springer, Berlin, pp. 187–192.

Barrufini, L., Garcia, P., Marini, A.J., Mota, B., Rocchini, P., 2000.

West Africa deep-water sedimentary processes and deposits: the

use of advanced imaging techniques in their recognition and

classification. 4th colloquium on the stratigraphy and the

palaeogeography on the south Atlantic, GeoLuanda 2000 Inter-

national Conference, Luanda, Angola, May 21–24.

Beaubouef, R.T., Friedmann, S.J., 2000. High resolution seismic/

sequence stratigraphic framework for the evolution of Pleisto-

cene intra slope basins, western Gulf of Mexico: depositional

models and reservoir analogues. In: Weimer, P., Slatt, R.M.,

Coleman, J., Rosen, N.C., Nelson, H., Bouma, A.H., Styzen,

M.J., Lawrence, D.T. (Eds.), Gulf Coast Section Society of

Economic Paleontologists and Mineralogists Foundation, 20th

Annual Bob F. Perkins Research Conference Proceedings,

Houston, Texas, December 3–6th, Deep-Water Reservoirs of

the World, pp. 40–60.

Belderson, R.H., Kenyon, N.H., 1976. Long-range views of sub-

marine canyons. Mar. Geol. 22, 69–74.

Belderson, R.H., Stride, A.H., 1969. The shape of submarine can-

yon heads revealed by Asdic. Deep-Sea Res 16, 103–104.

Bouma, A.H., Stelting, C.E., Coleman, J.M., 1985. Mississippi

Fan, Gulf of Mexico. In: Bouma, A.H., Normark, W.R., Barnes,

N.E. (Eds.), Submarine Fans and Related Turbidite Systems.

Springer-Verlag, New York, pp. 143–150.

Bouma, A.H., Coleman, J.M., Stelting, C.E., Kohl, B., 1989. Influ-

ence of relative sea level changes on the construction of the

Mississippi Fan. Geo. Mar. Lett. 9, 161–170.

Campion, K.M., Sprague, A.R., Mohrig, D., Lovell, R.W., Drze-

wiecki, P.A., Sullivan, M.D., Ardill, J.A., Jensen, G.N., Sicka-

foose, D.K., 2000. Outcrop expression of confined channel

complexes. In: Weimer, P., Slatt, R.M., Coleman, J., Rosen,

N.C., Nelson, H., Bouma, A.H., Styzen, M.J., Lawrence, D.T.

(Eds.), Gulf Coast Section Society of Economic Paleon-

tologists and Mineralogists Foundation, 20th Annual Bob

F. Perkins Research Conference Proceedings, Houston,

Texas, December 36th, Deep-Water Reservoirs of the World,

pp. 127–151.

Carter, R.M., 1988. The nature and evolution of deep-sea channel

systems. Basin Res. 1, 41–54.

Clifton, H.E., 1981. Submarine canyon deposits, Point Lobos,

California. In: Frizelli, V. (Ed.), Upper Cretaceous and Paleo-

cene Turbidites, Guide Book to Field Trip, vol. 6. Pacific

Section, S.E.P.M., Central California Coast, pp. 79–92.

Clifton, H.E., 1984. Sedimentation units in stratified deep-water

conglomerate, Paleocene submarine canyon fill, Point Lobos,

California. In: Koster, E.H., Steele, R.J. (Eds.), Sedimentolo-

gy of Gravels andConglomerates, Mem.-Can. Soc. Pet. Geol.,

vol. 10, pp. 429–441.

Cronin, B.T., 1994. Channel-fill architecture in deep-water

sequences: variability, quantification and applications. Unpub-

lished PhD thesis, University of Wales, 332 pp.

Cronin, B.T., 1995. Structurally-controlled deep-sea channel

courses: examples from the Miocene of South-east Spain and

the Alboran Sea, South-west Mediterranean. In: Hartley, A.J.,

Prosser, D.J. (Eds.), Characterization of Deep Marine Clastic

Systems, Special Publication-Geological Society of London,

vol. 94, pp. 113–133.


Cronin, B.T., Kidd, R.B., 1998. Heterogeneity and lithotype distri-

bution in ancient deep-sea canyons: Point Lobos deep-sea can-

yon as a reservoir analogue. Sediment. Geol. 115, 315–349.

Cronin, B.T., Kenyon, N.H., Woodside, J.M., Ivanov, M., den

Bezemer, T., Millington, J., Van Der Wal, A., Limonov, A.,

1995. Views of the Andarax submarine canyon: a meandering

system on an active tectonic margin. In: Pickering, K.T., Hiscott,

N., Smith, R., Kenyon, N.H. (Eds.), Atlas of Deep Water

EnvironmentsArchitectural Style in Turbidite Systems. Chap-

man and Hall, London, pp. 84–88.

Cronin, B.T., Celik, H., Hurst, A., Turkmen, I., in press. Mud prone

entrenched deep-water slope channel complexes from the Eo-

cene of eastern Turkey. In: Hodgson, D., Flint, S., Garfield, T.,

(Eds.), Submarine slope systems: processes, products and pre-

diction, Liverpool, Special Publication-Geological Society of

London, 54.

Damuth, J.E., Hayes, D.E., 1977. Echo character of the east Brazi-

lian continental margin and its relationship to sedimentary pro-

cesses. Mar. Geol. 24, 73–95.

Damuth, J.E., Kowsmann, R.O., Flood, R.D., Belderson, R.H.,

Gorini, M.A., 1983. Age relationships of distributary channels

on Amazon deep-sea fan: implications on fan growth pattern.

Geology 11, 470–473.

Deptuck, M.E., Steffens, G.S., Barton, M., Pirmez, C., 2003.

Architecture and evolution of upper fan channel-belts on the

Niger Delta slope and in the Arabian Sea. Mar. Pet. Geol. 20,

649–676.

Dore, A.G., Lundin, E.R., Birkeland, O., Eliassen, P.E., Fichler, C.,

1999. Principal tectonic events in the evolution of the northwest

European Atlantic Margin. In: Fleet, A.J., Boldy, S.A.R. (Eds.),

Petroleum Geology of NW Europe: Proceedings of the 5th

Conference. Geological Society of London, pp. 41–61.

Gardner, J.V., Kidd, R.B., 1983. Sedimentary processes on the

Iberian continental margin viewed by long-range side-scan

sonar: Part 1. Gulf of Cadiz. Oceanol Acta 6 (3), 245–254.

Gardner, J.V., Field, M.E., Lee, H., Edwards, B.E., Masson, D.G.,

Kenyon, N.H., Kidd, R.B., 1991. Ground-truthing 6.5-kHz side

scan sonographs: what are we really imaging?. J Geophys Res

96 (B4), 5955–5974.

Heezen B.C., Tharp M., 1961. Physiographic diagram of the South

Atlantic Ocean, the Caribbean Sea, the Scotia Sea, and the

eastern margin of the South Pacific Ocean (with explanation).

New York, Geological Society of America, scale 1:10,000,000.

Kenyon, N.H., 1986. Evidence from bedforms for a strong poleward

current along the upper continental slope of Northwest Europe.

Mar. Geol. 72 (1–2), 187–198.

Kenyon, N.H., Belderson, R.H., Stride, A.H., 1978. Channels,

canyons, and slump folds of the continental slope between

SW Ireland and Spain. Oceanol. Acta 1, 369–380.

Kenyon, N.H., Akhmetzhanov, A.M., Wheeler, A.J., van Weering,

T.C.E., de Haas, H., Ivanov, M.K., 2003. Giant carbonate mud

mounds in the southern Rockall Trough. Mar. Geol. 195, 5–30.

Le Danois, E., 1948. Les profoundeurs de la mer. Payot, Paris.

303 pp.

Le Pichon, X., Renard, V., 1982. Avalanching: a major process of

erosion and transport in deep-sea canyons: evidence from sub-

mersible and multi-narrow beam surveys. In: Scruton, R.A.,

Talwani, M. (Eds.), The Ocean Floor. John Wiley and Sons,

New York, pp. 113–128.

Malahoff, A., Embley, R.W., Fornari, D.J., 1982. Geomorphology

of Norfolk and Washington canyons and the surrounding con-

tinental slope and upper rise as observed from DSRVAlvin. In:

Scruton, R.A., Talwani, M. (Eds.), The Ocean Floor. John Wiley

and Sons, New York. 318 pp.

Mayall, M., Stewart, I., 2001. The architecture of turbidite slope

channels. In: Fraser, S.I., Johnson, H.D., Fraser, A.J., Evans,

A.M. (Eds.), Petroleum Geology of Deepwater Depositional

Systems—Advances in Understanding 3D Architecture. Geo-

logical Society, London.

Mayall, M., Syms, R., Jones, E., Henton, J., in press. Turbidite slope

channels—patterns of reservoir distribution and heterogeneity.

In: Hodgson, D., Flint, S., Garfield, T. (Eds.), Submarine Slope

Systems: Processes, Products and Prediction, Liverpool, Special

Publication-Geological Society of London, 54.

Morris, W.R., Busby-Spera, C., 1988. Sedimentologic evolution of

a submarine canyon in a forearc basin, Upper Cretaceous

Rosario Formation, San Carlos, Mexico. Bull. Am. Assoc.

Pet. Geol. 72, 717–737.

Mutti, E., 1985. Hecho turbdite system, Spain. In: Bouma, A.H.,

Normark, W.R., Barnes, N.E. (Eds.), Submarine Fans and

Related Turbidite Systems. Springer-Verlag, Inc., New York,

pp. 205–208.

Mutti, E., Seguret, M., Sgavetti, M., 1988. Sedimentation and

deformation in the Tertiary sequences of the Southern Pyrenees.

AAPG Mediterranean Basins Conference, Nice, France, Sep-

tember 1988, Field Trip Guide, vol. 7.

Naylor, D., Shannon, P.M., Murphy, N., 1999. Irish Rockall basin

region—a standard structural nomenclature. Petroleum Affairs

Division (Dublin, Rep. of Ireland), Special Publication 1/99.

Piper, J.W., Normark, W.R., 1971. Re-examination of a Miocene

deep-sea fan and fan-valley, Southern California. Geol. Soc.

Amer. Bull. 82, 1823–1830.

Posamentier, H.W., Vail, P.R., 1988. Eustatic controls on clastic

deposition II—sequence and systems tract models. In: Wilgus,

C.K., Hastings, B.S., Kendall, C.G.St.C., Posamentier, H.W.,

Ross, C.A., Van Wagoner, J.C. (Eds.), Sea-Level Changes—

An Integrated Approach, Special Publication of the Society of

Economic Palaeontologists and Mineralogists, vol. 42, pp. 125–

154. Tulsa.

Scruton, R.A., Talwani, M., 1982. The Ocean Floor. John Wiley and

Sons, New York. 318 pp.

Shepard, F.P., 1977. Geological Oceanography: Evolution of

Coasts, Continental Margins, and the Ocean Floor. Crane Rus-

sak and Co., New York.

Shepard, F.P., 1982. Submarine canyons: multiple causes and long-

time persistence. Am. Assoc. Pet. Geol. Bull. 65, 1062–1077.

Shepard, F.P., Emery, K.O., 1973. Congo submarine canyon and fan

valley. Am. Assoc. Pet. Geol. Bull. 57, 1679–1691.

Shepard, F.P., Marshall, N.F., 1978. Currents in submarine canyons

and other sea valleys. In: Stanley, D.J., Kelling, G. (Eds.),

Sedimentation in Submarine Canyons and Other Sea Valleys.

Hutchinson & Ross, Inc., Dowden, pp. 3–14.

Stoker, M.S., Hitchen, K., 2003. The rockall-porcupine margin. In:

Mienert, J., Weaver, P. (Eds.), European Margin Sediment Dy-


namics: Side-Scan Sonar and Seismic Images. Springer-Verlag,

Berlin, pp. 161–172.

Stoker, M.S., van Weering, T.C.E., Svaerdborg, T., 2001. A mid- to

late Cenozoic tectonostratigraphic framework for the Rockall

Trough. In: Shannon, P.M., Haughton, P.D.W., Corcoran, P.V.

(Eds.), The Petroleum Exploration of Ireland’s Offshore Basins,

Special Publication-Geological Society of London, vol. 188,

pp. 411–438.

Times Books, 1999. The Times Comprehensive Atlas of the World,

Millennium edition, plate 56. Times Books, London.

Twichell, D.C., Roberts, D.G., 1982. Morphology, distribution and

development of submarine canyons on the US Atlantic conti-

nental slope between Hudson and Baltimore. Geology 10,

408–412.

Weimer, P., 1990. Sequence stratigraphy, facies geometries, and

depositional history of the Mississippi Fan, Gulf of Mexico.

Bull. Am. Assoc. Pet. Geol. 74, 425–453.

Wonham, J.P., Jayr, S., Mougamba, R., Chuilon, P., 2000. 3D

sedimentary evolution of a canyon fill (Lower Miocene age)

from the Mandorove Formation, offshore Gabon. Mar. Pet.

Geol. 17, 175–197.

Wheeler, A.J., Kenyon, N.H., Ivanov, M.K., Beyer, A., Cronin,

B.T., McDonnell, A., Schenke, H.W., Akhmetzhanov, A.M.,

Satur, N., Zaragosi, S., 2003. Canyon heads and channel

architecture of the Gollum Channel, Porcupine Sea bight. In:

Mienert, J., Weaver, P. (Eds.), European Margin Sediment

Dynamics: Side-Scan Sonar and Seismic Images. Springer, Ber-

lin, pp. 183–186.



An evaluation of along- and down-slope sediment transport

processes between Goban Spur and Brenot Spur on

the Celtic Margin of the Bay of Biscay

M.J. Cunningham *, S. Hodgson, D.G. Masson, L.M. Parson

Challenger Division for Seafloor Processes, National Oceanography Centre, Empress Way, Southampton, SO14 3ZH, United Kingdom

Geological Survey of Ireland, Department of Communications, Marine and Natural Resources, Ireland

Petroleum Affairs Division, Department of Communications, Marine and Natural Resources, Ireland

BT, United Kingdom

Flag Telecom, United Kingdom

Gemini, United Kingdom

Tyco Telecommunications, United Kingdom

Global Crossing and Global Marine Systems, United Kingdom

Abstract

Multi-beam bathymetry and backscatter, 3.5 kHz pinger profiles, side-scan sonar and seabed samples have been examined to

evaluate along- and down-slope sedimentary processes along the Celtic Margin shelf and upper slope in water depths of 200 to

1500 m. The continental shelf and slope are indented and dissected by major canyon systems. The shelf is characterised by major

northeast – southwest trending sand banks that are orthogonal to the shelf edge. Along the shelf edge, several fields of asymmetric

sandwaves oriented orthogonal to the canyon axes indicate sediment transport into the canyon heads. Less commonly, sandwaves

with weak asymmetry suggest sediment transport onto the shelf. These may be reworked and are partly overprinted by more recent

sandwaves. Down-slope sediment transport by turbidity currents is the dominant process through the major canyons. Recent

faulting has also played a role in canyon development. Turbidity currents are most likely initiated by faulting, and/or slope failure of

the walls that bound the canyon head drainage basins and sediment migration from the shelf. This leads to deep incision of sinuous

thalwegs in the upper reaches of canyon floors and downcutting and sediment transport on the mid to lower continental slope. The

canyons are V-shaped in the upper reaches and become U-shaped progressively down-slope, suggesting they represent either a

transition from erosive to depositional processes or sediment bypass conduits carrying sediment between the shelf and abyssal

plain. Over-bank spill from canyons leads to deposition of unconfined turbidite deposits (muds) on the intervening canyon spurs.


Keywords: Sedimentary processes; Morphology; Celtic Margin; Multi-beam bathymetry; Sediment waves; Submarine canyons

0037-0738/$ - s

doi:10.1016/j.se

* Correspondi

Southampton, S

E-mail addre

79 (2005) 99–116


dgeo.2005.04.014

ng author. Navigation Safety Branch, Maritime and Coastguard Agency, Bay 2/30, Spring Place, 105 Commercial Road,

O15 1EG, UK. Tel.: +44 23 8032 9198.

ss: [email protected] (M.J. Cunningham).

Fig. 1. Regional setting of study area. a) Shaded bathymetric image

of SW Eire, SWApproaches and northern Bay of Biscay. Resolution

is 2 km. b) 3D image of Celtic Margin showing an amalgamation of

multi-beam bathymetry sourced from the Geological Survey of

Ireland, the Irish Petroleum Affairs Division of the Department of

Communications, Marine and Natural Resources (Ireland), subma-

rine telecommunications industry data and GEBCO. Horizonta

resolution is 150 m. Note that most prominent canyons along the

Celtic Margin shelf trend north-northeast – south-southwest and

then develop into a series of smaller canyons further west. The

small canyons however coalesce down-slope with the main canyon

system. BS - Brenot Spur, CSSB - Celtic Sea Sand Banks, GC -

Gollum Channel, GS - Goban Spur, GSDA - Great Sole Drainage

Area, KAC - King Arthur Canyon, LCB - La Chapelle Bank

LSDA - Little Sole Drainage Area, MT - Meriadzek Terrace, PB -

Porcupine Bank, SS - Shamrock System, WS - Whittard System

M.J. Cunningham et al. / Sedimentary Geology 179 (2005) 99–116100

1. Introduction

Over the last 10 to 15 years, multi-beam (swath)

bathymetric sonar systems have become the dominant

survey instrumentation at all water depths: deep ocean

(e.g., Exon et al., 1997), continental slope (e.g., Shaw

and Courtney, 1997; Todd et al., 1999; McAdoo et al.,

2000), continental shelf, shallow bays and estuaries

(Courtney and Fader, 1994). When integrated with

other information such as seabed samples and seismic

profiles, it is possible to chart bathymetry, and study

morphology and geology in unprecedented detail in-

cluding the mapping of fault scarps, sediment slides,

debris flows and turbidite deposits, the recognition of

sediment transport pathways to offshore depocentres,

studies of fauna and habitat, and the evaluation of

sediment transport processes.

The study area is located on the Celtic margin of

the Bay of Biscay between 11819VN/48855VW and

9825VN/48830VW and 150 to 4500 m water depths

(Fig. 1). Whilst new multi-beam bathymetric coverage

reveals exceptional details of sea floor morphology

within the study area, our knowledge of sedimentary

processes is poorly constrained. For example, which

canyons contribute to sediment flux between the shelf

and the abyssal plain? Is there a genetic link between

the Celtic Sea Sand Banks located on the continental

shelf as a sediment source (Marsset et al., 1999), and

the sandy sediments of the Celtic Deep Sea Fan

(Zaragosi et al., 2000) at the foot of the continental

rise? What sediment patterns and processes affect

individual channels: which are likely to be experien-

cing erosion and canyon head retreat, which are swept

clean or filling up with sediment? To improve our

knowledge of the Celtic Margin, we need to better

understand processes such as bottom currents and

their temporal variability, the structural and erosional

characteristics of canyon evolution, bedload transport,

and sedimentary depositional patterns by tidal cur-

rents. There is also a significant commercial aspect

(e.g., for the submarine telecommunications industry)

to increasing our understanding of sedimentary pro-

cesses such as the assessment of risk from slope

instability and the prevention of cable breaking.

The main aims of this paper are to gain a better

understanding of sedimentation processes on the Cel-

tic Margin through the interpretation of integrated

multi-beam bathymetry/backscatter, side-scan sonar,

l

,

.

3.5 kHz profiles, bottom sediment grabs and gravity

cores. We first outline the regional setting of the study

area, describe our observations of erosive and depo-

sitional features, and then discuss the significance and

main implications for along- and down-slope sedi-

ment transport processes.

The data presented in this paper are based on 1)

submarine telecommunications industry cable route

surveys, 2) multi-beam bathymetry/backscatter and

3.5 kHz profiles of the Geological Survey of Ireland

M.J. Cunningham et al. / Sedimentary Geology 179 (2005) 99–116 101

(GSI) collected as part of the Irish National Seabed

Survey (Cullen, 2003), and 3) multi-beam bathymetry

of the Irish Petroleum Affairs Division of the De-

partment of Communications, Marine and Natural

Resources, Ireland; acquired in 1996 as part of the

United Nations Convention Law Of the Sea (Part VI,

Article 76).

2. Materials and methods

The bathymetry/backscatter and shallow seismic

data presented in this paper were collected onboard

the RV Bligh in 2001 using Kongsberg Simrad

EM1002S (water depths b1000 m) and EM120

(water depths N1000 m) echo sounders and a hull-

mounted 3.5 kHz Oretech 3010 S transceiver, respec-

tively. Seismic reflectors were mapped from 3.5 kHz

profiles and subsequently transformed into contoured

3.5 kHz echofacies maps of penetration depth. Pene-

tration contours were further constrained using shaded

relief bathymetry by forcing the interpolation algo-

rithms to follow bathymetric contours and avoid areas

of steep slope (N158).Data supplied by the submarine telecommunica-

tions industry were acquired onboard the NO Jean

Charcot in 2000. These data include multi-beam ba-

thymetry/backscatter, shallow seismic and side-scan

sonar and were collected using a Konsberg Simrad

EM120 echo sounder, a hull-mounted 3.5 kHz ORE

4�4 Pinger and a GeoAcoustics Deepwater Model

159 fish, respectively. Bathymetry and backscatter

were processed and gridded using open source MB-

System software (Caress and Chayes, 1995). Shallow

seismic sub-bottom profiles (SBP) and side-scan

sonar (SSS) were supplied as analogue rolls. In addi-

tion to the geophysical/morphological data, a gravity

corer with a 3 m long barrel and a Shipek grab

sampler were used to acquire seabed samples during

survey operations (Fig. 1b).

3. Regional setting

The Celtic Margin is a sediment-starved passive

margin trending west-northwest – east-southeast (Fig.

1; Roberts et al., 1981; de Graciansky et al., 1985).

The margin is characterised by a relatively steep

continental slope with a mean gradient of 118, butlocally attaining very steep gradients to vertical along

canyon walls. The geological evolution of the margin

is a combination of several major phases of tectonic

activity (de Graciansky et al., 1985) associated with

the Palaeozoic formation and subsequent Mesozoic

disintegration of Pangea, and the initiation of North

Atlantic rifting and sea-floor spreading (Ziegler,

1981). The Celtic Margin shelf is characterised by a

series of fault-bounded, rift basins trending west-

southwest – east-northeast (Dingle and Scrutton,

1979; Roberts et al., 1981; Ziegler, 1987). Seawards

of the margin, there is a series of tilted and rotated

fault blocks bounded by north-northwest – south-

southeast trending, listric normal faults (Dingle and

Scrutton, 1979; Roberts et al., 1981; de Graciansky et

al., 1985), which relate to rifting prior to seafloor

spreading. Late Cretaceous eustatic sea-level rise

combined with a cessation of tectono-sedimentary

activity resulted in the progressive overstepping of

earlier basin margins and the deposition of a wide-

spread transgressive sequence (Roberts et al., 1981;

Ziegler, 1987).

Post-Pliocene, there was a widespread marine

transgression due to continuing subsidence (Naylor

and Shannon, 1982) with relatively low sedimentation

rates (Dingle and Scrutton, 1979). It has been sug-

gested that rejuvenated Late Neogene –Holocene can-

yon cutting and slope wasting has removed much of

the earlier Neogene sediments deposited on the West-

ern Approaches slope (Evans et al., 1990).

3.1. Hydrodynamics

The hydrodynamics of the Celtic Margin are typi-

fied by high-energy conditions, in particular related to

storm surges and spring tides, which transport sediment

from the near-shore to the shelf-edge (Reynaud et al.,

1999; Zaragosi et al., 2000). Sediment transport may

also occur in the form of mass flows, either due to

erosion of the Celtic Sea Sand Banks, or from

regressive canyon head erosion (Reynaud et al.,

1999; Zaragosi et al., 2000).

Tidal conditions in the Celtic Sea are typically

vigorous and sediments are transported south-west-

ward (Zaragosi et al., 2000), related to ebb-dominated

tidal flow (Reynaud et al., 1999). These currents

decrease in strength from the southeast (around La


Chapelle Bank, velocities close to the seabed are

locally 0.9 m s�1, Heathershaw et al., 1987) to the

northwest (currents of 0.2 m s�1 are more typical of

the shelf southeast of the Goban Spur, Huthnance et

al., 2001). At the shelf-edge, the dynamics of these

large tidal currents generate correspondingly large

internal tides, which may also be important for sedi-

ment transport across the slope area (Reynaud et al.,

1999; Wollast and Chou, 2001; Huthnance et al.,

2001). In addition, internal waves may serve to rein-

force bottom current velocities, which cause sediment

erosion (Huthnance et al., 2001) and can occur at

depths of up to 400 – 500 m on the Celtic Margin

(Pingree and Le Cann, 1989).

Whilst bottom current data in the region are rela-

tively sparse, results from the Ocean Margins EX-

change project (Wollast and Chou, 2001) show

typical along-slope residual current velocities, close

to the shelf-break, of 0.05 m s�1 (Huthnance et al.,

2001). Such currents are responsible for north-west-

ward transport and are associated with a general

northerly flow west of the UK driven by the North

Atlantic Current and at depth, by the Mediterranean

Outflow Water (900 – 1000 m) and the Lower Deep

Water (N2500 m) (Colling, 2001; Huthnance at al.,

2001). Internal tides are probably important for trans-

port and reworking of fine- to very fine-grained sedi-

ments within canyons and on the intervening spurs

(see Section 4.4).

3.2. Submarine canyons

Steep-sided canyons crosscut the continental shelf-

edge and deeply incise the margin (Fig. 1). The

orientation of canyons at the shelf edge is normally

north-northwest – south-southeast and north-north-

east – south-southwest, although numerous smaller

canyon segments also occur and vary towards east –

west trends. The shape and form of the canyon heads

is semi-circular with a concave profile. The canyons

provide conduits for the transport of sediment from

the shelf to the abyssal plain (e.g., Zaragosi et al.,

2000) and for over-bank turbidity currents, which

deposit on the intervening terraces and spurs.

A number of the shelf-break canyon heads may

possibly be the seaward expression of Pleistocene

river mouths, which supplied large volumes of sed-

iment directly to the shelf edge (Hadley, 1964;

Marsset et al., 1999; Droz et al., 2003; Bourillet

and Lericolais, 2003). The present-day deposits of

the area are dominated by marine bioclastic sands

(Bouysse et al., 1979), with fine-grained material in

the form of terriginous silts and clays, which by-

pass the hydrodynamically energetic shelf to settle

on the continental slope below water depths of

500 m (Auffret et al., 1979).

4. Morphology and sedimentary processes —Celtic

Margin

The Celtic Margin shelf/slope transition has a gen-

eral west-northwest – east-southeast trend from the

King Arthur Canyon (eastern margin of the Goban

Spur) to the Brenot Spur (Fig. 1a). The slope break

occurs at water depths of between 170 and 300 m,

landward of which the shelf morphology (at a cell

resolution of 25 m) is relatively smooth (Fig. 1b). The

margin is dissected by canyons with dominant south-

southwest and south-southeast trends. In general, two

types occur along the margin: 1) canyons with rela-

tively long, narrow upper reaches and V-shaped pro-

files that incise the shelf-break; and 2) canyons with

relatively short, broad upper reaches, U-shaped pro-

files and heads deeper than the shelf-break on the

continental slope. This pattern breaks down further

west where morphology consists of smaller canyon

segments in various orientations, but is re-established

towards the King Arthur Canyon (Fig. 1a,b).

Morphological evidence for sediment transport is

seen throughout the canyon systems. Some of the

most important features are described in the following

Sections, beginning at the shelf edge and progressing

down-slope.

4.1. Sediment waves north of Brenot Spur

The Brenot Spur forms an important divide between

two main canyon networks, the Whittard and Sham-

rock systems. The Whittard system is located to the

west of the spur and drains the Great Sole drainage area

(linked to the southern end of the lowstand Irish Sea

river system) along the Celtic Margin. To the east, the

Shamrock system drains the Little Sole drainage, which

is primarily associated with the lowstand English

Channel river system (Bourillet and Lericolais, 2003;


Zaragosi et al., 2003). The western margin of the

Brenot Spur is bounded by a major north – south trend-

ing canyon. This canyon swings to a north-northwest –

south-southeast trend upslope and offsets three north-

northeast – south-southwest canyons (Fig. 1b).

The largest of the canyons occurs in the east. Two

sets of sediment waves occur on either side of the

canyon head, one to the southeast and one to the

northwest (Figs. 1 and 2). Pinger (3.5 kHz) profiles

show zero penetration of the sediments within the

waves, which suggests that the substrate is sandy

and that these bedforms are sandwaves.

The sandwaves southeast of the main canyon head

are oriented northwest – southeast (Set A in Fig. 2)

and cover a minimum area of 150 km2 (sandwaves

extend beyond our bathymetric coverage). They are

Fig. 2. Sandwaves. Shaded bathymetric image of two sets of prominent sed

inferred. West-northwest striping in this and subsequent images represents

consistently asymmetrical with angular crests and lee

slopes facing southwest (Fig. 2) and are mainly par-

allel. They have wavelengths of 450 – 650 m, and

typical relief of 3 – 5 m. These sandwaves occur in

water depths of 170 – 200 m and continue into the

canyon head area (Fig. 2).

The larger set of sandwaves (covering an area

N200 km2 in water depths of 180 to 230 m) lies

some 3 km to the northwest of the canyon head and

is oriented west-northwest – east-southeast (Set B in

Fig. 2). The morphology of these sandwaves is quite

different to that of Set A. They are sub-parallel,

swinging clockwise to a more northwest–southeast

trend towards the west, and show frequent bifurcation

(Fig. 2). They have a lower slope angle and opposite

sense of asymmetry (lee slopes facing northeast) com-

iment waves from which opposing sediment transport directions are

artefacts (ship-tracks).


pared to Set A (Fig. 3). Some are also near symmet-

rical and tend to have more rounded crests and

troughs in comparison to Set A. Wavelengths vary

between 350 and 650 m with relief of 2 – 10 m. A

secondary field of east–west trending sandwaves

occurs on the northeast slopes of the main sandwaves.

These secondary features have a wavelength of 250 –

350 m, and relief of 1 – 3 m (Fig. 2). The wavelength

of Set B sandwaves decreases westward to 250 – 300

m. The decreasing relief and subsequent disappear-

ance of the sandwaves are associated with a change of

slope aspect (Fig. 3), which marks the uppermost

extent of another canyon head.

To the west, there is a series of arcuate sandwaves

at the head of the dominant north-northwest – south-

southeast canyon, covering an area of N10 km2 (Fig.

4). The sandwaves become slightly sinuous towards

the northeast and have an overall northwest – southeast

trend. These sandwaves have a relief of 0.5 – 2 m and a

wavelength of 100 – 120 m. They show minor asym-

metry, generally towards the southwest, but occa-

sionally to the northeast.

In summary, the crests of the sandwaves are or-

thogonal to the axes of canyon heads. The strong

asymmetry of Set A shows sediment movement is

towards the canyon heads and illustrates that the

Fig. 3. Slope-aspect of sandwaves. a) Gradient map (averaged over 75 m

head. b) Aspect image showing facing directions of both sets of sandwaves

images represent flat areas. Note the change in aspect related to the cany

canyons have a sediment source. The potential build

up and subsequent failure of sandwaves in the canyon

heads may initiate gravity and/or turbidity currents, an

important agent in canyon incision and erosion. The

sandwaves of Set B show subtle asymmetry, are N1

km distance from the canyon heads, and give an

inferred sediment movement away from the canyon

heads. These features can be explained by considering

the bifurcation and rounded morphology of the sand-

waves. The presence of a secondary set of sandwaves

suggests that Set B originally formed during the last

glacial low-stand, possibly through wave-dominated

processes and the secondary set are forming under

present day current regimes. This is in harmony with

the primary sandwaves now being re-worked and

more symmetrical in form. It is also possible, how-

ever, that these sandwaves are related to temporal

variations in storm events, which reflect the prevailing

southwesterly trade winds.

4.2. Canyon head- faulting, slumping

An important feature of the north-northwest –

south-southeast trending canyon are the well-devel-

oped drainage basins (also known as bamphitheatre

rimsQ e.g., Belderson and Kenyon, 1976) (Fig. 4) on

length scale). Note bathymetric Profile 1 is upslope from a canyon

and canyon heads. Areas of grey on image and subsequent daspectTon head to the south.

Fig. 4. Tectonics and sediment erosion. a) Shaded bathymetric image at head of north-northwest – south-southeast canyon. The eastern margin

of the canyon is fault controlled with a minimum vertical displacement of 120 m. Most recent movement post-dates major canyon scarp retreat

of a tributary system. b) 3D shaded relief showing east-southeast trending hanging wall tributary entering the main canyon system on the

western margin. Also shown are sediment waves, which infer a sediment transport direction towards the southwest.


the eastern margin (marked as AT on Fig. 4). Within

the catchment boundary there is a well-developed

channel network with interfluves bounded to the

west by a scarp forming the main canyon margin.

The headwalls of the canyon occur at water depths

of 200 – 330 m and 200 – 460 m on the west and

eastern margins, respectively (Fig. 4).

Erosion and downcutting of the canyon floor have

resulted in the carving of a thalweg, with the floor

having an average down-slope gradient of 38 com-


pared to average gradients of 0.68 on the shelf and 118on the canyon walls (Fig. 5). The main canyon has a

single tributary on the western margin, which trends

west-northwest – east-southeast and is 6.5 km in

length. This canyon tributary has a thalweg with

moderate sinuosity of 1.17 compared to the main

canyon thalweg sinuosity of 1.08 (Schumm and

Khan, 1972). There is also deep incision in the trib-

utary canyon floor of up to 150 m (Fig. 4) and a

channel gradient of 68, which is coupled with strongly

developed incised meanders.

On the northern margin of the tributary, above its

confluence with the main canyon channel, there are

two small plateaux (b2 km2 in total), one of which

dips gently west-northwest, the other dipping gently

east-southeast (Fig. 4). There is no obvious drainage

system crossing either of these platforms and they

may represent rotational slides from over-steepened

canyon walls, or terraces that may have formed by

lateral migration of the thalweg (e.g., see Mulder et

al., 2004).

4.2.1. Upper reaches

The upper reaches of the canyon head are desig-

nated as those above the confluence of the main

channel and its primary tributary (Fig. 4). Here the

canyon is oriented approximately north – south and

the main canyon thalweg shows a low degree of

sinuosity (1.01) with weakly developed meanders

Fig. 5. Slope-aspect of faulted canyon. a) Gradient map (averaged over 60

on the eastern margin of the canyon. This fault trace continues across par

prominent sediment waves to the northeast margin of the canyon head.

and a down-slope gradient of 38. The upper reaches

are characterised by a more distinct incision of tribu-

tary canyon heads on the eastern margin of the main

channel than those on the western margin.

4.2.2. Lower reaches

At the channel confluence there is a subtle change

in the orientation of the canyon axis, towards the

north-northwest – south-southeast. Below this point,

the canyon is essentially linear in character and has

an average down-slope gradient of 28. The east-north-east edge of this part of the canyon is defined by a

steep scarp with an average height of 130 – 140 m and

a gradient in excess of 608 (Fig. 4). The top of the

scarp lies at water depths of 420 – 460 m and is linear

in plan. It can be traced into the upper reaches of the

canyon (though less well defined than in the lower

reaches), over a distance of approximately 7 km and

has a south-southwest facing slope. In general, this

scarp is parallel to the lower reaches of the canyon

floor.

Given the linear nature, steepness and relatively

large relief of the eastern canyon margin scarp, we

interpret this feature as a fault. To the northwest the

scarp disappears, although it may coincide with the

beginning of channel incision in the upper reaches of

the canyon. This suggests that there may be a signi-

ficant tectonic control on down-slope erosion and

subsequent sediment deposition.

m length scale). Note the steep northwest – southeast trending fault

t of the amphitheatre. b) Aspect image of canyon system. Note the


The fault offsets the drainage basins (bamphitheatre

rimsQ) and with sufficient time, displacement along

the fault could result in a steepening of channel

gradient from the drainage basin. The timing of

such changes in fluvial systems is poorly understood;

similarly there are no documented, quantified studies

of the response of submarine systems to such tec-

tonic movements. However, it is clear that insuffi-

cient time has passed since fault movement for these

changes to occur, as the drainage basins do not

appear to have developed a more bmatureQ morphol-

ogy, e.g. rounded interfluves, gentler slopes, up-

stream migration of channel knickpoints (currently

located on the fault scarp), and formation of a par-

abolic thalweg (re-equilibrium with post-fault dis-

placement boundary conditions). Therefore, we

Fig. 6. Canyon incision and slumping. a) Shaded relief image showing thr

displays deep incision of up to 150 m in its upper reaches, with a V-shaped

U-shaped profile (Profile 2). The Eastern canyon is broader in form with

Central canyon forms a broad, flat plateau. The white star represents the lo

60 m length scale). Note very high angle slopes clearly define the incised c

map: this shows that shelf surfaces (interfluves) are relatively flat (shaded g

artefacts. CC - Central canyon, DB - Drainage basin, DIC - Deeply incise

speculate that this fault is an active structure (see

Discussion).

4.3. Down-slope sediment transport

The main canyon tributary of the fault-controlled

canyon is separated by a small ridge from a broad

U-shaped canyon further west. The orientation of this

canyon is northeast – southwest and forms the eastern

component of a set of three sub-parallel canyons

(Fig. 6). The two other canyons, referred to here as

the Central and Western canyons, are much narrower

and more V-shaped in their upper reaches. At the

head of the western canyon, a series of asymmetric

northwest – southeast trending bedforms indicate a

south-southwest sediment transport direction.

ee major north-northeast trending canyons. The canyon to the west

profile (Profile 1). Compare this to further down-slope where it has a

a far lower degree of incision. The interfluve between this and the

cation of a gravity core. b) Image of maximum slope (averaged over

hannel and its terraces, whilst the shelf is relatively flat. c) Gradient

rey). Closely spaced stripes (parallel to depth contours) are gridding

d canyon, EC - Eastern canyon, WC - Western canyon.


4.3.1. Eastern canyon

The morphology of the canyon head is less incised

than those further west and occurs on the slope at a

depth of 400 m. This may suggest that it is younger

than the shelf-indenting Central and Western canyons

(Fig. 6) where retrogressive mass wasting of slope

sediments along headwalls (bamphitheatre rimsQ) hasled to headward migration and eventual indention of

the shelf (e.g., Twichell and Roberts, 1982; Farre et

al., 1983). However, it has also been suggested that

canyons are eroded by turbidity currents on the upper

slope/shelf edge (Daly, 1936). Slope failures lead to

gravity flows moving down-slope through pre-exist-

ing bathymetric lows and thus lead to canyon evolu-

tion. Subsequent flows widen the canyons by

retrogressive canyon wall failure, i.e. undercutting of

the canyon walls leading to instability and thus caus-

ing slope failure (Pratson et al., 1994; Pratson and

Coakley, 1996).

A gravity core on the drainage divide between the

Eastern canyon and the canyon tributary to the east

consists of green to brown homogeneous soft, slight-

ly clayey, silty, fine sand. The substrate material is

capable, given the existence of appropriate flow

regimes (possibly internal tides), of producing close-

ly spaced sandwaves. However, the apparent lack of

bedforms and the broad U-shape of this canyon head

might indicate that the system once extended further

upslope and has subsequently been cut by the fault-

controlled canyon system to the east, and, as shown

in Section 4.4, erosive processes are active in this

canyon.

4.3.2. Central canyon

The Central canyon occurs in water depths of

190 – 240 m, with retrogressive canyon wall failure

occurring on both margins. The canyon thalweg

shows moderate sinuosity (1.23) and has a V-shaped

profile (Profile 2, Fig. 6). The canyon has an average

down-slope channel gradient of 2.58 and typical can-

yon wall gradients of 10 – 148 (Fig. 6b). Drainage

basin development is more extensive on the eastern

margin (DB on Fig. 6). Whilst it is clear that slumping

has occurred on the eastern margin, the basin now

appears to be infilling with sediment. In general, this

drainage basin, with an ideal parabolic thalweg, is

more maturely developed than the fault-related drain-

age basin (Fig. 4).

4.3.3. Western canyon

The Western canyon has a typical down-slope

channel gradient of 38, with average canyon wall

gradients in the order of 11 – 158. Canyon margins

occur in water depths of 200 – 300 m and the canyon

has a deeply incised channel with drainage basins on

the margins. The drainage basins are less well devel-

oped than those of the Central canyon. The primary

features of the deeply incised thalweg include two sets

of parallel terraces and steep channel walls of 458(Fig. 6a,b). Consequently the canyon has a stepped,

relatively narrow, V-shaped profile (Profile 1, Fig. 6),

which broadens downstream to a U-shape profile

(Profile 2, Fig. 6). The channel has a relatively linear

morphology and is 100 – 150 m deep. Terraces are

more strongly developed on the eastern margin of the

canyon (Fig. 6b), suggesting an increased rate of

incision and channel migration to the west.

4.3.4. Interfluve

The interfluve between the Western and Central

canyons has a width of 2 km at the canyon heads,

but broadens down-slope to 3.5 km at its widest

point. This plateau lies in water depths of ~200 m

and is typified by a low down-slope gradient (typi-

cally 1 – 28) towards the south-southwest. The inter-

fluve probably represents relict continental shelf

bounded to the east and west by the two canyon

systems (Profile 2, Fig. 6).

4.4. Relict canyons

Down-slope and to the east of the Eastern can-

yon, a broad terrace forms average slope gradients of

1 – 2.58, in water depths of 680 – 1200 m. This ter-

race is crosscut by three sub-parallel canyons (Fig.

7), trending west-northwest – east-southeast in their

upper reaches, and swinging north – south down-

slope. These crosscutting canyons occur in water

depths of 700 – 1100 m and their morphology is

atypical with respect to other Celtic Margin canyon

systems. They are U-shaped in cross-section, with

broad, sub-horizontal floors 1 – 1.5 km across and

300 – 350 m deep. Average down-slope canyon floor

gradients are in the region of 3 – 4.58. There is no

obvious incision of the canyon walls or canyon

floors. This suggests that these canyons represent

areas of sediment deposition.

Fig. 7. Sediment deposition and relict canyons. a) 3D Shaded relief image showing a spur bounded on the west by the Eastern canyon, and

cut by broad U-shaped canyons (Profile 1). West of the flow divide, these canyons are convex upward in longitudinal profile and to the east,

they are concave upward (Profiles 2 and 3). Towards the south, the spur steepens and has a convex upward slope; long wavelength, low

amplitude sediment waves are apparent (Profile 4). b) Bathymetric image of the same area overlain with an interpolated map of sediment

penetration based on 3.5 kHz profiles. Note the areas of high (light) and low (dark) penetration on the terrace surfaces and canyon floors,

respectively. c) 3D shaded multi-beam backscatter image showing areas of high backscatter in the canyon floors and low backscatter on the

terraces. CF - Canyon floor.


At the headward end of the canyons there is a

north – south drainage divide (Fig. 7). East of the

divide, canyon floors slope towards the west-south-

west (with spurs sloping mainly south) and steepen

towards the divide (canyon floor gradient V 68)forming a lip on the eastern margin of the divide

(Fig. 8). West of the divide, canyon floors slope

north-northeast at a more typical gradient of 3 – 4.58

Fig. 8. Sediment erosion and overspill deposition. Shaded relief image of transverse (relict) canyons and surrounding area. Side-scan sonar

images show that erosive processes appear to be dominant in the Eastern canyon and that sediment waves occur in the floors of the canyons.

CF - Canyon floor, D - Divide, GC - Gravity core.


until they are captured by the walls of the neigh-

bouring canyon system. Spurs slope predominantly

southwest.

It is unclear if headward erosion of the canyons

has led to a progressive narrowing of the divide, or

whether they represent an older trend that is now

truncated by younger canyons where turbidity over-

flow has resulted in sediment deposition on the spurs,

with partial burial of the canyons. A study of canyons

to the east of New Jersey (western North Atlantic) has

shown similar patterns where abandoned, and partial

to fully buried, canyons are offset by modern canyons

(Pratson et al., 1994; Pratson and Coakley, 1996).

The canyon floors are associated with areas of low

3.5 kHz seismic penetration and high 95 kHz back-

scatter (Fig. 7b,c) which may imply basement subcrop

or relatively coarse sand. A gravity core from the edge

of the most northerly canyon head and a bottom

sediment grab from between the two southern canyons

(Fig. 7) consist of brown-grey, sandy, silty clay with

occasional shell fragments and correspond with high

penetration and low backscatter. In the canyon heads

to the east of the divide, the side-scan sonar shows a

series of relatively closely spaced sediment waves

(Fig. 8). The sediment waves are generally orthogonal

to the canyon axes and show an east – southeast sed-


iment transport direction. However, these sediment

waves are too closely spaced to be resolved by

multi-beam bathymetry (50 m cell resolution). To

the west of the divide side-scan sonar shows that the

Eastern canyon is swept clean of sediment with ex-

tensive areas of bedrock exposure (Fig. 8).

Side-scan sonar data show that the terrace surface

between the active and relict canyons is generally

featureless. However, to the southwest where the

terrace slope gradient increases to form a convex

northeast – southwest profile, (Profile 4, Fig. 7) there

is a series of relatively high relief, long wavelength

sediment waves (Fig. 8). The sediment waves have a

wavelength of 1500 – 1800 m and relief of ~30 m.

These features are approximately parallel to each

other, broadly follow the slope contours and have an

overall northwest – southeast orientation. The sedi-

ment wave crests are generally orthogonal to the

upslope canyon axes and coincide with areas of high

3.5 kHz seismic penetration (~10 to N80 m). In cross-

sectional profile they form a rounded, terraced mor-

phology where the sub-horizontal slopes have a gra-

dient of b18 towards the northeast and the steep

slopes face southwest with a gradient of 8 – 118(Fig. 8). 3.5 kHz seismic profiles reveal that these

sediment waves have a greater depositional thickness

on the sub-horizontal slopes facing the sediment trans-

port direction.

In summary, the high backscatter intensities and

low 3.5 kHz penetration of the canyon floors cou-

pled with the existence of closely spaced sediment

waves is opposite in character to the intervening

canyon spurs with no detectable sediment waves.

This shows that relatively coarse material is accumu-

lating in the canyons with fines being preserved on

the spurs, and may imply that along-slope processes

affect these relict canyons, with only the coarsest

material being preserved.

Toward the south of the relict canyons the terrace

surface is covered by a series of low relief, long

wavelength sediment waves which coincide with

high 3.5 kHz seismic penetration, locally in excess

of 80 m. Internally, these sediment waves have a

greater depositional thickness on the sub-horizontal

lee slopes, which face up-slope. These bedforms have

been detected and described in other areas (e.g., Wynn

and Stow, 2002) and are believed to have been de-

posited by unconfined turbidity currents.

5. Discussion

5.1. Sediment supply and transport

On the Celtic Margin, at the shelf/slope transi-

tion, sandwave fields occur immediately upslope of

major canyon heads (Fig. 9). The sediment wave

crests are mainly perpendicular to the canyon axes,

with lee slopes facing towards the canyon heads.

This suggests that sediment transport is towards the

canyon systems. In general, the sandwaves lie to the

northeast of the canyon heads and this, combined

with the direction of asymmetry of Set A sandwaves

(Fig. 2), ties in with the regional southwesterly active

sediment transport direction (Heathershaw et al.,

1987).

A relatively large field of sandwaves to the

northwest of the Brenot Spur (Set B, Fig. 2) is

more equivocal in its association with present-day

sediment transport into the canyon system. These

sandwaves are more rounded in morphology, and

where asymmetry exists the inferred sediment trans-

port direction is towards the northeast. It is there-

fore suggested that these bedforms represent relict

features that originally developed during the last

sea-level lowstand under a current regime different

to that of the present day, and are now being

reworked to equilibrate with the existing current

regime.

The source of the shelf edge sandwaves is probably

to the north-east where the tidally dominated Celtic

Sea area is characterised by a series of elongate, linear

sand banks with a maximum length of 200 km, width

of 7 km, and height of up to 60 m (Reynaud et al.,

1999). The sand bank crests are oriented broadly

orthogonal to the shelf edge and are at an average

water depth of 150 m (Fig. 1a, Berne et al., 1998;

Marsset et al., 1999). The origin of these sand banks is

uncertain.

Previous workers have suggested that the Celtic

Sea Sand Banks may be tidal, formed during the last

glacial sea-level lowstand when tidal velocities were

high enough to generate them (e.g. Pantin and Evans,

1984; Berne et al., 1998; Zaragosi et al., 2000).

Alternatively, the sand banks may be the remnants

of a shallow marine delta channel-levee system

formed at the mouth of the English Channel River

(Marsset et al., 1999; Droz et al., 2003).

Fig. 9. Inferred sediment transport direction on shelf-break. Northeast shaded bathymetric image from ~150 to 400 m bathymetric contour

showing shelf-break and heads of major north-northeast-south-southwest trending canyons. The data is artificially cut off at 400 m to capture the

canyon heads and associated sandwaves. There is an obvious association of sediment waves at major canyon heads, with inferred direction of

sediment transport being normal to the canyon axes. In most cases, sediment transport is towards the shelf-break (canyon heads), but there is

also a series of sub-symmetrical to weakly asymmetrical sandwaves showing an inferred transport direction to the northeast. Given their

morphology (see Section 4.1), these are probably in the process of being reworked and originally formed during the last glacial low-stand. CC -

Central canyon, EC - Eastern canyon, FCC - Fault controlled canyon, WC - Western canyon.


5.2. Structure

The character of the Celtic Margin has been

strongly influenced by seafloor spreading and the

opening of the North Atlantic Ocean (at around 53

Ma), Palaeocene – Eocene Alpine foreland deforma-

tion, and mid-Cenozoic Pyrenean orogenesis (Zieg-

ler, 1987; Naylor and Shannon, 1982; Tucker and

Arter, 1987; Naylor, 2001; Bourillet and Lericolais,

2003). These events have resulted in major fault

blocks bounded by north-northwest – south-southeast

trending, listric normal faults (de Graciansky et al.,

1985).

About 20 km west-northwest of the Brenot Spur,

there is a canyon of north-northwest – south-southeast

orientation (see Section 4.2). On the eastern margin of

the canyon, there is a steep, linear scarp with a relief

of 130 – 140 m. There is also a drainage basin on the

eastern margin, which is clearly offset by the fault and

where insufficient time has passed for the basin to re-

equilibrate with its boundary conditions. These obser-

vations strongly suggest that the canyon is actively

fault controlled (Fig. 4).

The age of the fault is difficult to constrain. The

last regional compressional event associated with

Alpine foreland deformation occurred between the

Oligocene and Miocene (Ziegler, 1987; Cook,

1987) and it is therefore likely that the fault is at

least of similar age. However, the present-day fault

scarp relief of N100 m and the fact that movement

post-dates the morphologically fresh drainage basin

on the eastern canyon margin, suggests a much

younger age. Supporting evidence comes from the

magnitude of downcutting and incision that has oc-

curred down-slope of this canyon (Fig. 10). The

trend of the fault is consistent with other relatively

young, north-northwest – south-southeast trending

faults known from around the British Isles, which

include the Neogene Sticklepath fault of SW Eng-

land (Bristow and Hughes, 1971) and the Codling–

Newrey faults of the Irish Sea (e.g., Geoffroy et al.,

1996).

Fig. 10. 3D shaded relief of the entire Celtic Margin from the shelf to the foot of the continental rise. Arrows show down-slope flow direction

based on slope aspect. BS - Brenot Spur, CC - Central canyon, EC - Eastern canyon, FCC - Fault controlled canyon, RC - Relict canyons, SW -

Sediment waves, WC - Western canyon.


The interaction of sedimentary processes and

structure within the canyon is difficult to quantify,

particularly in the upper reaches where back-wall

retrogressive mass erosion of the eastern margin is

more developed when compared to the western mar-

gin. This greater development may also be, in part, a

function of bottom current movement along the Celt-

ic Margin, which would bring more sediment to the

eastern margin.

5.3. Upper to middle slope sediment deposition

In cross-sectional profile, the canyons of the Celtic

Margin are primarily V-shaped at the shelf edge,

becoming increasingly flat-bottomed to U-shaped

down-slope. The Western canyon (Fig. 6) is a good

example of this change in morphology since it is

strongly V-shaped in its upper reaches close to the

shelf-break, but becomes progressively more U-

shaped with increasing depth. A first-order interpre-

tation of this pattern is erosion dominating the upper

reaches and deposition on the lower reaches. How-

ever, it may also be that canyon morphology and

sediment flows are presently in a state of equilibrium,

and that the canyons act as bypass conduits carrying

sediment to the Celtic Sea Deep Fan which is pres-

ently dominated by low-density, muddy turbidity cur-

rent deposits (Zaragosi et al., 2000, 2003). This would

imply that flows are of insufficient size to overspill the

canyon system and that the deposition of fines on

spurs is probably related to background, terrestrially

derived sediment.

Spurs vary from being sharp-crested to relatively

flat in transverse profile, but are usually convex up-

ward in longitudinal profile and in some cases are

crossed by relict canyons that have been abandoned

and subsequently incised by newer canyon systems.

The Eastern canyon appears to be less incised than

the Central and Western canyons. This could be due to

differences in current velocities, sediment supply and/

or substrate. Alternatively, it could be related to the

development of the fault controlled canyon system to

the east, which has captured the sediment supply that

once fed this canyon.

The terrace east of the Eastern canyon is crossed

by three U-shaped canyons, with a drainage divide

near its western margin (Fig. 7). The multi-beam

bathymetry, backscatter and lack of 3.5 kHz pene-

tration imply relatively coarse material on the canyon

floors, suggesting that they are swept clean of finer

material. The U-shaped profile of these canyons

combined with the occurrence of bedforms on the

canyon floors (close to the drainage divide) supports

this proposal.

The unconfined turbidite deposits forming sedi-

ment waves towards the southern end of the terrace


are likely to have originated upslope in the canyon

heads. There is ample evidence from faulting,

drainage basins (bamphitheatre rimsQ), deep incision

and changes of sinuosity (particularly incised mean-

ders) for the retrogressive mass wasting and inden-

tion of the canyon heads, leading to indention of

the shelf-break. Over-steepening of the canyon

headwalls and drainage basins will lead to slope

failure, which will initiate coarse-grained gravity

and turbidity currents that downcut and deepen

the canyon system. The high rates of incision at

canyon headwalls, particularly the Western canyon

(Fig. 6), show that there may also have been a

recent influx of coarse-grained sediment (probably

at the end of the last glaciation), leading to frequent

turbidity currents (although the possibility of slow

rates of incision that have been ongoing over long

periods of time cannot be discounted). Down-slope,

the canyon walls will subsequently over-steepen

and fail, leading to a widening of canyons. There

is little deposition in the main canyons, although

sediment waves, which occur on many interfluve

areas, suggest that down-slope turbidity currents

spill out of the canyon system and form over-

bank deposits on the terrace/spur surfaces (Fig. 10).

Hence, we suggest that recent mud-rich flows have

been of sufficient size to overspill the canyon system

leading to fine sediment deposition on the continental

slope.

Our observations are in agreement with previous

studies (Zaragosi et al., 2000, 2003). For example, the

Celtic Deep Sea Fan, which spreads out at the foot of

the continental rise at water depths of 4200 – 4900 m

(Zaragosi et al., 2003), is a mature, mud-rich system

(Zaragosi et al., 2000; Droz et al., 2003) with a

depositional history commencing in the Miocene

(Droz et al., 1999). Isotope studies show that much

of the fan material has been sourced from high-density

turbidity currents (forming sand-rich deposits) during

glacial low-stands, whereas at present, fan sediments

are derived from the outer shelf by low-density turbi-

dite currents (forming mud-rich deposits) (Zaragosi et

al., 2000).

The Celtic Sea Sand Banks on the continental shelf

are the obvious source for fine muds reaching the fan

system, whilst very fine sands, silts and clays are

deposited further upslope on the spurs as canyon

overspill unconfined turbidite deposits.

6. Conclusions

Multi-beam bathymetry and backscatter, 3.5 kHz

pinger profiles, side-scan sonar and seabed samples

have been integrated to evaluate along- and down-

slope sedimentary processes along the Celtic Margin.

The main conclusions are:

1) Asymmetrical sandwaves occur along the Celtic

Margin shelf-break, are orthogonal to the canyon

axes, and are inferred to demonstrate sediment

transport into the canyon heads.

2) Near symmetrical sandwaves show well-rounded

crests with no obvious association to canyon heads

and occur along the Celtic Margin shelf-break.

Where asymmetry exists, the inferred sediment

transport direction is away from the shelf edge.

3) These dsymmetricalT sandwaves probably formed

prior to the end of last glacial low-stand at ~14 ka

BP. The smaller set of superimposed sandwaves

indicates progressive reworking through present-

day bottom current activity. Alternatively, the sym-

metrical sandwaves may reflect temporal variations

in north-eastward moving storm events.

4) Active down-slope sediment transport in the form

of turbidity currents is the dominant process in the

upper reaches of canyons characterised by deep

incision of sinuous thalwegs into the canyon floors.

5) Drainage basins occur along the margins of the

canyon heads and are interpreted to result from

retrogressive mass wasting and indention of the

Celtic Margin shelf.

6) Active faulting appears to control the development

of a north-northwest – south-southeast canyon. The

canyon shows rapid and deep down-cutting of the

continental shelf down-slope of the fault and off-

sets earlier canyon systems.

7) Earlier west-northwest – east-southeast canyons lo-

cated mid-slope are now areas of sediment depo-

sition. Along-slope sediment transport of fines

occurs in relict canyons that are sub-parallel to

the shelf-break. This process has led to the forma-

tion of closely spaced sediment waves.

8) Active canyons are dominated by sediment trans-

port in the mid to lower slope, but over-bank spill

of turbidity currents has led to the deposition of

muds and clays. Further down-slope, these form

unconfined turbidite deposits.


9) Canyons are V-shaped in the upper reaches and

become U-shaped progressively down-slope. This

may indicate a transition from erosive to deposi-

tional down-slope processes, or that the canyons

act as bypass conduits carrying sediment to the

abyssal plain.

From this work, it appears that the Celtic Sea

Sand Banks of the Celtic Margin have a genetic

link to the Celtic Deep Sea Fan at the foot of the

continental rise. The main agent of sediment trans-

port in the canyon heads is by slope failure and the

seaward migration of sandwaves whose sediment is

sourced from the Celtic Sea Sand Banks. These will

initiate gravity currents, which down-slope, will in-

cise, deepen and widen the canyons. However, over-

bank sediment deposition leads to along-slope depo-

sition in relict canyons and on spurs and terraces,

which further down-slope form unconfined turbidite

deposits.

Acknowledgements

This study was made possible through funds pro-

vided by NERC (grant ner/t/s/2000/01013). We

would also like to thank our project partners, BT,

Flag Telecom, Gemini, Global Crossing Systems,

Global Marine Systems and Tyco Telecommunica-

tions for allowing us access to their submarine cable

data. We thank Fugro Survey and Svitzer for unarch-

iving and supplying us with certain cable route

survey data. We wish to thank Sean Cullen, Archie

Donovan, Mick Geoghagan and Roger Sweetman of

the Geological Survey of Ireland for allowing us to

use data from the Irish National Seabed Survey and

Dr. Peter Croker of the Petroleum Affairs Division of

the Department of Communications, Marine and

Natural Resources, Ireland, for allowing us to use

multi-beam bathymetric data. We benefited from

reviews by R. Brunt and S. H. Chough and fruitful

discussions with Dr Russell Wynn.

References

Auffret, G.A., Pastouret, L., Cassat, G., De Charpel, O., Gravatte, J.,

Guennoc, P., 1979. Dredged rocks from the Armorican and

Celtic margins. In: Montadert, L., Roberts, D.G., Auffret,

G.A., et al., (Eds.), Initial Reports on the Deep Sea Drilling

Project, vol. 48. US Government Printing Office, Washington

DC, pp. 995–1013.

Belderson, R.H., Kenyon, N.H., 1976. Long-range sonar views of

submarine canyons. Mar. Geol. 22, 69–74.

Berne, S., Lericolais, G., Marsset, T., Bourillet, J.-F., de Batist, M.,

1998. Erosional offshore sand ridges and lowstand shorefaces:

examples from tide- and wave-dominated environments of

France. J. Sediment. Res. 68 (4), 540–555.

Bourillet, J.-F., Lericolais, G., 2003. Morphology and seismic stra-

tigraphy of the manche paleoriver system, western approaches.

In: Mienert, J., Weaver, P. (Eds.), European Margin Sediment

Dynamics: Side-Scan Sonar and Seismic Images. Springer-Ver-

lag, Berlin, pp. 229–232.

Bouysse, P., Lelann, F., Scolari, G., 1979. Les sediments super-

ficiels des approches occidentales de la Manche. Mar. Geol. 29,

107–135.

Bristow, C.M., Hughes, D.E., 1971. A Tertiary thrust fault on the

southern margin of the Bovey Basin. Geol. Mag. 108, 61–68.

Caress, D.W., Chayes, D.N., 1995. New software for proces-

sing sidescan data from sidescan-capable multibeam sonars.

In: Wernli, R. (Ed.), Oceans 95 MTS/IEEE: Challenges of our

Changing Global Environment, Conference Proceedings, vol. 2.

Marine Technology Society, Washington DC, pp. 997–1000.

Colling, A., 2001. Ocean Circulation, 2nd edition. Butterworth-

Heinemann, Oxford, p. 286.

Cook, D.R., 1987. The Goban Spur — exploration in a deep-water

frontier basin. In: Brooks, J., Glennie, K. (Eds.), Petroleum

Geology of North-West Europe. Graham and Trotman Ltd.,

London, pp. 623–632.

Courtney, R.C., Fader, G.B.J., 1994. A new understanding of the

ocean floor through mulitbeam mapping. Science Review.

Bedford Institute of Oceanography, Dartmouth, Nova Scotia,

pp. 9–14.

Cullen, S., 2003. Irish national seabed survey — an introductory

overview. Hydro. J. 109, 22–25.

Daly, R.A., 1936. Origin of submarine bcanyonsQ. Am. J. Sci., Ser. 5

(31), 401–420.

de Graciansky, P.C., Poag, W.C., Cunningham, R., 1985. The

Goban Spur transect: geologic evolution of a sediment-starved

passive continental margin. Bull. Geol. Soc. Am. 96, 58–76.

Dingle, R.V., Scrutton, R.A., 1979. Sedimentary succession and

tectonic history of a marginal plateau. (Goban Spur, Southwest

of Ireland.). Mar. Geol. 33, 45–69.

Droz, L., Auffret, G.A., Savoye, B., Bourillet, J.-F., 1999. The

Celtic deep-sea fan: stratigraphy and sedimentary evolution.

C. R. Acad. Sci., IIa 328 (3), 173–180.

Droz, L., Auffret, G.A., Savoye, B., 2003. The Celtic deep-sea fan:

seismic facies, architecture and stratigraphy. In: Mienert, J.,

Weaver, P. (Eds.), European Margin Sediment Dynamics:

Side-Scan Sonar and Seismic Images. Springer-Verlag, Berlin,

pp. 233–238.

Evans, C.D.R., Hillis, R.R., Gatliff, R.W., Day, G.A., Edwards,

J.W.F., 1990. The geology of the Western English channel and

its Western approaches. United Kingdom Offshore Regional

Report. HMSO, London, p. 93.


Exon, N.F., Moore, A.M.G., Hill, P.J., 1997. Geological framework

of the South Tasman Rise, South Tasmania, and its sedimentary

basins. Aust. J. Earth Sci. 44, 561–577.

Farre, J.A., McGregor, B.A., Ryan, W.B.F., Robb, J.M., 1983.

Breaching the shelfbreak; passage from youthful to mature

phase in submarine canyon evolution. In: Stanley, D.J.,

Moore, G.T. (Eds.), The Shelfbreak: Critical Interface on Con-

tinental Margins, Special Publication, vol. 33. Society of Eco-

nomic Paleontologists and Mineralogists, Tulsa, pp. 25–39.

Geoffroy, L., Bergerat, F., Angelier, J., 1996. Brittle tectonism in

relation to the Palaeogene evolution of the Thulean/NE Atlantic

domain: a study in Ulster. Geol. J. 31, 259–269.

Hadley, M.L., 1964. Wave induced bottom currents in the Celtic

Sea. Mar. Geol. 2, 164–167.

Heathershaw, A.D., New, A.L., Edwards, P.D., 1987. Internal tides

and sediment transport at the shelf break in the Celtic Sea. Cont.

Shelf Res. 7 (5), 485–517.

Huthnance, J.M., Coelho, H., Griffiths, C.R., Knight, P.J., Rees,

A.P., Sinha, B., Vangriesheim, A., White, M., Chatwin, P.G.,

2001. Physical structures, advection and mixing in the region of

Goban Spur. Deep-sea Res., Part 2, Top. Stud. Oceanogr. (48),

2979–3021.

Marsset, T., Tessier, B., Reynaud, J.-Y., de Batist, M., Plagnol,

C., 1999. The Celtic sea banks: an example of sand body

analysis from very high-resolution seismic data. Mar. Geol.

158, 89–109.

McAdoo, B.G., Pratson, L.F., Orange, D.L., 2000. Submarine land-

slide geomorphology, US continental slope. Mar. Geol. 169,

103–136.

Mulder, T., Cirac, P., Gaudin, M., Bourillet, J.-F., Tranier, J.,

Normand, A., Weber, O., Griboulard, R., Jouanneau, J.-M.,

Anschutz, P., Jorissen, F.J., 2004. Understanding continent-

ocean sediment transfer. EOS 85, 257–262.

Naylor, D., 2001. Geology of offshore Ireland. In: Holland,

C.H. (Ed.), The Geology of Ireland. Dunedin Academic Press,

Edinburgh, pp. 443–491.

Naylor, D., Shannon, P.M., 1982. The Geology of Offshore Ireland

and West Britain. Graham and Trotman, London, pp. 161.

Pantin, H.M., Evans, C.D.R., 1984. The Quaternary history of the

central and southwestern Celtic Sea. Mar. Geol. 57, 259–293.

Pingree, R.D., Le Cann, B., 1989. Celtic and Armorican slope and

shelf residual currents. Prog. Oceanogr. 23, 303–338.

Pratson, L.F., Coakley, B.J., 1996. A model for the headward

erosion of submarine canyons induced by downslope-eroding

sediment flows. GSA Bull. 108, 225–234.

Pratson, L.F., Ryan, W.B.F., Mountain, G.S., Twichell, D.C., 1994.

Submarine canyon initiation by downslope-eroding sediment

flows: evidence in late Cenozoic strata on the New Jersey

continental slope. GSA Bull. 106, 395–412.

Reynaud, J.-Y., Tessier, B., Berne, S., Chamley, H., de Batist, M.,

1999. Tide and wave dynamics on a sand bank from the deep

shelf of the Western channel approaches. Mar. Geol. 161,

339–359.

Roberts, D.G., Masson, D.G., Montadert, L., de Charpal, O., 1981.

Continental margin from the Porcupine Seabight to the Armor-

ican marginal basin. In: Illing, L.V., Hobson, G.D. (Eds.),

Petroleum Geology of the Continental Shelf of North-West

Europe. Heyden and Son Ltd., London, pp. 455–473.

Schumm, S.A., Khan, H.R., 1972. Experimental study of channel

patterns. Nature 233, 407–409.

Shaw, J., Courtney, R.C., 1997. Multibeam bathymetry of glaciated

terrain off southwest Newfoundland. Mar. Geol. 143, 125–135.

Todd, B.J., Fader, G.B.J., Courtney, R.C., Pickrill, R.A., 1999.

Quaternary geology and surficial sediment processes, Browns

Bank, Scotian Shelf, based on multibeam bathymetry. Mar.

Geol. 162, 165–214.

Tucker, R.M., Arter, G., 1987. The tectonic evolution of the North

Celtic Sea and Cardigan Bay basins with special reference to

basin inversion. Tectonophysics 1–4 (137), 291–307.

Twichell, D.C., Roberts, D.G., 1982. Morphology, distribution, and

development of submarine canyons on the United States Atlan-

tic continental slope between Hudson and Baltimore Canyons.

Geol. 10, 408–412.

Wollast, R., Chou, L., 2001. Ocean margin EXchange in the North-

ern Gulf of Biscay: OMEX I. An introduction. Deep-sea res.,

Part 2, Top. Stud. Oceanogr. 48, 2971–2978.

Wynn, R.B., Stow, D.A.V., 2002. Classification and characterization

of deep-water sediment waves. Mar. Geol. 192, 7–32.

Zaragosi, S., Auffret, G.A., Faugeres, J.C., Garlan, T., Pujol, C.,

Cortijo, E., 2000. Physiography and recent sediment distribution

of the Celtic Deep-sea Fan, Bay of Biscay. Mar. Geol. 169,

207–237.

Zaragosi, S., Auffret, G.A., Voisset, M., Garlan, T., 2003. Mor-

phology and depositional processes of the Celtic Fan, Bay of

Biscay. In: Mienert, J., Weaver, P. (Eds.), European Margin

Sediment Dynamics: Side-Scan Sonar and Seismic Images.

Springer-Verlag, Berlin, pp. 239–243.

Ziegler, P.A., 1981. Evolution of sedimentary basins in North-West

Europe. In: Illing, L.V., Hobson, G.D. (Eds.), Petroleum Geo-

logy of the Continental Shelf of North-West Europe. Heyden

and Son Ltd., London, pp. 3–39.

Ziegler, P.A., 1987. Celtic Sea –Western approaches area: an over-

view. Tectonophysics 137 (1–4), 285–289.



Contrasting styles of shelf sediment transport and deposition in a

ramp margin setting related to relative sea-level change

and basin floor topography, Turonian (Cretaceous)

Western Interior of central Utah, USA

Chris M. Edwards*, David M. Hodgson, Stephen S. Flint, John A. Howell 1

Stratigraphy Group, Department of Earth Sciences, University of Liverpool, 4 Brownlow Street, Liverpool L69 3GP, United Kingdom


Abstract

The Turonian lower Ferron Sandstone and Juana Lopez Members of the Mancos Shale in the Western Interior foreland basin

form a series of regressive shoreface to shelf depositional complexes sourced from the Sevier fold-thrust belt to the west.

Continuous, dip-parallel outcrop exposures around the northern and eastern edges of the Tertiary San Rafael Anticline (SRA) in

central Utah provide an insight into the origins of anomalous, marine mudstone-encased sandstones.

Lower Ferron strata consist of both coarsening-and-thickening upward and sharp-based shoreface successions, interpreted to

have been deposited as sequential highstand and falling stage systems tracts. Overlying these shallow marine deposits is a

surface that marks sediment bypass that is correlated down-dip into coarse-grained, cross-bedded sandstones, interpreted as the

deposits of channelised turbidity flows. These turbidites are considered to be the products of sustained hyperpycnal flows

exiting rejuvenated lowstand rivers based on sustained flow indicators such as dune-scale cross-bedding, absence of genetically

related slumps and delta front deposits in up-dip areas, surfaces of sediment bypass, coarser grain-sizes than the underlying

shoreface deposits, and regional palaeogeography. The flows are interpreted to have been ignitive in proximal areas down a

tectonically-induced slope in the uplifted area of the Farnham Dome. The change in depositional style from falling-stage to

lowstand systems tracts is attributed to increased rates of sediment supply to the shoreline, promoting hyperpycnal flows. A

flooding surface caps the interval and separates the lower Ferron Sandstone from the overlying Juana Lopez Member. This

interval in contrast, consists only of heterolithic, parallel and ripple-laminated graded sandstone–mudstone couplets interpreted

to have been deposited by turbulent, storm-induced geostrophic flows. The couplets are arranged into bundles and are

interpreted as parasequences, which in turn are arranged into a progradational parasequence set deposited during a period of

sea-level highstand.

0037-0738/$ - s

doi:10.1016/j.se

* Correspondi

E-mail addre1 Present addr

Norway.

9 (2005) 117–152


dgeo.2005.04.011

ng author. Present address: ExxonMobil Exploration Company, 233 Benmar, Houston, Texas 77060, USA.

ss: [email protected] (C.M. Edwards).

ess: Department of Earth Sciences/Centre for Integrated Petroleum Research, University of Bergen, Allegt. 41, N-5007 Bergen,

C.M. Edwards et al. / Sedimentary Geology 179 (2005) 117–152118

Divergent sole mark and bedform palaeocurrents from both members suggest slight eastward deflection of the turbulent

flows off the proto-SRA. Further evidence for this topography is inferred from an absence of turbidites along the western side of

the structure. The inferred influence of local basin floor structures on sedimentation of these units has important implications for

structural models for this basin, for generalized sequence stratigraphic models applied to ramp-type foreland basin sequences

and for elucidating the origin of anomalous marine mudstone-encased sandstone bodies.


Keywords: Western Interior Basin; Turbidites; Hyperpycnal flows; Geostrophic flows; Sequence stratigraphy

1. Introduction

The Western Interior foreland basin formed as a

flexural response to Late Cretaceous thrust sheet load-

ing in the Sevier thrust belt to the west (Jordan, 1981;

Wiltschko and Dorr, 1983; DeCelles et al., 1995). The

combined effects of subsidence and high Cretaceous

eustatic sea-level (Haq et al., 1988) created an epicon-

tinental seaway that stretched over 5000 km from

Arctic Canada in the north to the Gulf of Mexico in

the south. Shorelines periodically advanced and

retreated into the basin in tandem with changes in

relative sea-level and sediment supply (McGookey et

al., 1972; Williams and Stelck, 1975; Kauffman,

1977). Basin physiography was represented by a shal-

low, eastward-dipping ramp and probable water depths

of no more than a few hundred ms (Kauffman, 1977).

In the central Utah sector of this basin, Late Cretaceous

palaeogeography was dominated by eastward prograd-

ing shorelines fed by rivers that transported sediment

from the uplifted Sevier orogen to eastward-tapering

shorelines in the Castle Valley area. These sediments

are now superbly exposed in the badland and mesa

topography of the Uinta and Piceance basins of central

Utah and western Colorado (Fig. 1).

Within this basin, a number of Late Cretaceous,

marine mudstone-encased sandbodies are found phys-

ically disconnected from their contemporaneous shor-

elines by distances of many tens to hundreds of

kiloms. The depositional origins of these enigmatic

units have been the focus of intense debate in recent

years where the absence of a regional physiographic

shelf-edge precludes the application of conventional

deep-water sediment distribution models (c.f. Posa-

mentier et al., 1991; Kolla and Perlmutter, 1993).

Upper Cretaceous isolated sandbodies of this foreland

basin, such as the Cardium Formation of Alberta (e.g.,

Bergman and Walker, 1987; Plint, 1988; Walker and

Plint, 1992) and the Shannon Sandstone of Wyoming

(e.g., Tillman and Martinsen, 1987; Walker and Berg-

man, 1993; Bergman, 1994), commonly show an

alignment parallel to the palaeoshoreline and consist

of upward-coarsening successions ornamented by a

variety of wave-and storm-generated bedforms. Early

explanations argued their origins as the products of

migrating, linear doffshore barsT. This interpretation

was largely based on observations from modern shelf

environments (e.g., Swift and Rice, 1984; Tillman and

Martinsen, 1987). However it was subsequently con-

sidered that this model could not adequately explain

long transport pathways, the transport of granular

sediment or the coarsening upward motif associated

with these genetic units (Walker, 1984; Plint, 1988).

The resultant explanation relied on changes in relative

sea-level to displace shorefaces long distances into the

basin during relative sea-level fall before isolating the

dislocated shoreface with a sea-level rise (e.g., Berg-

man and Walker, 1987; Plint, 1988; Walker and Plint,

1992; Walker and Bergman, 1993; Bergman, 1994).

In the Utah sector of the Western Interior Basin,

some isolated sandstones have been attributed to rela-

tive sea-level fluctuations (e.g., Taylor and Lovell,

1995; Van Wagoner, 1995; Hampson et al., 1999).

Thin, mud-rich sandstones of the Campanian Prairie

Canyon Member of the Mancos Shale have been inter-

preted as the products of up-dip palaeovalley incision

into sand-rich highstand shorefaces of the Blackhawk

Formation with coeval down-dip forced regression of

muddy shorelines (Hampson et al., 1999). Similar en-

visaged sea-level fluctuations and incised valley devel-

opment have been applied in the interpretation of the

Tocito sandbodies in the northern New Mexico seg-

ment of the basin (Jennette and Jones, 1995). An

alternative hypothesis used to explain the occurrence

of channelised deposits such as the Prairie Canyon

Member adopts an offshore-directed delta plume

Fig. 1. Distribution of the lower Ferron Sandstone and Juana Lopez Members of central Utah, southwestern United States. The interval forms a

westward thickening wedge towards a high-subsidence zone adjacent to the Sevier Orogen. (A) Sediments sourced from the orogen were shed

into the epicontinental Western Interior Seaway that stretched from northern Canada to the Gulf of Mexico (Modified from Kauffman, 1984 and

Eaton and Nations, 1991). (B) To the north lie the Uinta Mountains and, at the time of lower Ferron Sandstone deposition, the bVernal HighQ ofRyer and Lovekin (1986). (C) Middle Turonian strata rim the northern edge of the San Rafael Anticline and extend southwards and eastwards

along the foot of the Book Cliffs. To the west, the lower Ferron is the primary coal bed methane reservoir horizon in the Drunkards Wash and

Helper Fields sited within the Ferron Coal Trend (from Montgomery et al., 2001). Structural elements are taken from Witkind (1988).

C.M. Edwards et al. / Sedimentary Geology 179 (2005) 117–152 119

model associated with periods of sea-level highstand

(Swift et al., 1987; Chan et al., 1991; Cole and Young,

1991; Cole et al., 1997; Stevens and Chaiwongsaen,

2003; Patterson, 2005). This model assumes fluvial

effluent exiting the delta front to have a bulk density

greater than the ambient density of the marine water

into which it enters in order to create a hyperpycnal

underflow (Bates, 1953). A characteristic inverse to

normal grading of individual event beds as a result of

the waxing and waning flood hydrograph (e.g., Mulder

et al., 1998) have been reported in sandstones of the

Prairie Canyon succession (Stevens and Chaiwong-

saen, 2003; Patterson, 2005). The interpretation of

these thin bedded sandstone intervals as hyperpycnal

turbidites provides a further mechanism for the depo-

sition of basinal isolated sandstones.

Fig. 2. Chronostratigraphy of the lower Ferron and Juana Lopez

stratigraphy along Castle Valley. The lower Ferron Sandstone

(dVernal deltaT) is an older and northerly counterpart to the upper

Turonian upper Ferron Sandstone (dLast Chance deltaT) that was

sourced from an area to the southwest of southern Castle Valley

Ammonite biostratigraphy is from Kauffman et al. (1993). Termi-

nology of depositional units defined by Cotter (1975) is given in

small type.


This paper introduces an unusual succession that

encompasses well-exposed bstrandedQ sandstones be-longing to the Turonian lower Ferron Sandstone, with

contemporaneous shoreline strata located at higher

levels on the shelf. The studied section of the outcrop

extends from a road cut near the town of Wellington,

Utah to the town of Green River on Interstate High-

way 70, a down-dip distance of over 70 km and

permits the correlation of shallow marine strata with

shoreline-detached sandstones. The shallow marine

succession provides a reliable sequence stratigraphic

framework into which the deposition of the mudstone-

encased sandstones may be placed. Overlying depos-

its of the Juana Lopez Member are equally well

exposed, although this interval cannot be traced up-

dip into an equivalent succession of shoreface strata.

Nonetheless, their relative sea-level context can be

sufficiently established using existing regional corre-

lations. Given the reliance upon understanding rela-

tive sea-level as a control on the deposition of isolated

shelf sandstones, this information is critical to their

correct interpretation. In addition to the relative sea-

level understanding, this paper presents data that im-

plicate the role of intra-basinal structures in the depo-

sition of the lower Ferron and Juana Lopez Members.

1.1. Geologic context

The lithostratigraphic term dFerron SandstoneTrefers to Turonian-age fluvio–deltaic units extending

from southern Castle Valley to the westernmost Book

Cliffs in the area of Price, Utah (Lupton, 1916; Spie-

ker and Reeside, 1925; Hale, 1972). These deposi-

tional wedges thicken westwards in the direction of

the thrust belt and taper eastwards into the Tununk

Shale of the Mancos Shale (Fig. 2). Deposits of this

age record the earliest major shoreline regression into

the Western Interior Seaway in central Utah and are

internally characterized by higher frequency sea-level

cycles. One northeastward-prograding depositional

system was sourced from an area southwest of Castle

Valley, and a second, separate system in northern

Castle Valley delivered sediment from an area north-

west of Price to a shoreline located at the northern-

most tip of the San Rafael Anticline (Fig. 2). These

two separate successions constitute the informally

known dLast ChanceT (southern) and dVernalT (north-ern) deltas (Hale, 1972; Ryer, 1991). Detailed sedi-

.

mentological and stratigraphic studies have revealed

that the Vernal and Last Chance deltas were not

contemporaneous. The Vernal delta deposits constitute

an older and entirely separate regressive wedge from

the younger Last Chance delta complex (Cotter, 1975;

Ryer and McPhillips, 1983). This relationship was

confirmed by macrofossil dating using the ammonite

biostratigraphic framework of the Western Interior

Seaway where the lower part of the Ferron Sandstone

was determined to lie within the upper middle Tur-

onian biozone of Prionocyclus hyatti (Molenaar and

Cobban, 1991). The northern succession was infor-

mally renamed the dlower Ferron SandstoneT to estab-

lish a clear distinction from the dupper Ferron

SandstoneT, which refers to the deposits of the Last

Chance delta outcropping in southern Castle Valley

(Ryer and McPhillips, 1983). This paper focuses on

the deposits and stratigraphy of the Vernal delta or

lower Ferron Sandstone, the older and northern coun-

terpart to the well-known upper Ferron Sandstone.

The lower Ferron Sandstone was first described in

detail by Cotter (1975) who subdivided it into the

dClawsonT, dFarnhamT and dWashboard UnitsT. He

noted that sediment was sourced from the west and

northwest and delivered to a NE–SW oriented shore-

line that tapered southwards and eastwards to its

down-dip termination in the Tununk Shale (Fig. 2).

The Clawson Unit was considered to be a shallow

marine shoreface deposit that crops out west of the

San Rafael Anticline (SRA) and the younger Wash-

board and Farnham Units were ascribed to shelf and

tidal inlet depositional environments, respectively


(Cotter, 1975). Ryer (1981) subsequently interpreted

the lower Ferron Sandstone as a storm-and wave-

dominated shoreline that was sourced from an area

to the north. Several authors have proposed the exis-

tence of an unconformity spanning early Cenomanian

to late middle Turonian, related to the uplift of an

intra-basinal culmination and possible source area in

the area of the Uinta Mountains (e.g., Weimer, 1962;

the dVernal HighT of Ryer and Lovekin, 1986; Fig. 1).

Merewether and Cobban (1986) and Molenaar and

Wilson (1990) concluded that the area of uplift was

centered in northwestern Colorado, southwestern

Wyoming and parts of northeastern Utah. The pro-

posed culmination was considered to be part of an

array of active mid-Cretaceous basin structures in the

Western Interior Seaway at this time (Merewether and

Cobban, 1986).

Riemersma and Chan (1991) and Gardner (1995)

attempted to place the lower Ferron Sandstone in a

relative sea-level context. Thick sandstones of the

Farnham Unit described by Cotter (1975) were related

to a relative sea-level fall in the Farnham Dome area

(Molenaar and Cobban, 1991; Riemersma and Chan,

1991; Gardner, 1995). The lower Ferron Sandstone of

central Utah constitutes part of a period of a wide-

spread regional regression across the southwestern

part of the Western Interior Basin. It is biostratigra-

phically constrained to be equivalent to shoreface

successions of the dRegressive Coastal SandstoneT ofthe Frontier Formation of the southern Uinta Moun-

tains of northeastern Utah (Hale and Van de Graaff,

1964; Molenaar and Wilson, 1990), an unconformity

at the base of the Dry Hollow Member in the western

Uinta Mountains (Ryer, 1977; Molenaar and Wilson,

1990) and to the dFerron SandstoneT of the Henry

Mountains (Peterson et al., 1980).

Less well understood are the overlying deposits of

the Juana Lopez Member, which crop out along the

base of the western Book Cliffs. Molenaar and Cob-

ban (1991) interpreted the thin mudstone and sand-

stone interbeds of this unit were derived from a source

area to the north and may have been deposited in an

outer shelf environment by storm-generated currents.

In a regional context, deposition of the Juana Lopez

Member was coeval with northeastward progradation

of the upper Ferron Sandstone along the western flank

of the San Rafael Anticline in southern Castle Valley

(Gardner, 1995) and incision into underlying regres-

sive shoreface sandstones of the Frontier Formation

on the south side of the Vernal High (Molenaar and

Wilson, 1990). At present however, no detailed anal-

ysis of depositional mechanisms or stratigraphic ar-

chitecture of this succession exist. Furthermore, its

stratigraphic relationship to the lower Ferron Sand-

stone has not been addressed.

The San Rafael Anticline around which the lower

Ferron and Juana Lopez deposits are exposed (e.g.,

Fig. 3), is traditionally interpreted as a latest Creta-

ceous to early Tertiary feature (Lawton, 1983; Law-

ton, 1986; Witkind, 1988; Franczyk and Pitman,

1991). However, speculations on pre-Tertiary tectonic

activity in this area have been well documented (e.g.,

Peterson, 1986; Eaton et al., 1990; Molenaar and

Cobban, 1991; Gardner, 1995; Martinson et al.,

1998). Peterson (1986) suggested that the SRA is a

remnant late Paleozoic structure that was activated

during the Early Jurassic to early Middle Jurassic,

based on isopach maps and distribution of these age

strata. Eaton et al. (1990) provided evidence of tec-

tonic activity from the identification of reworked

exotic clasts within the Mancos Shale that could

only be derived from older rocks deposited prior to

marine flooding of the Western Interior Seaway pro-

viding supporting evidence to observation by Mole-

naar and Cobban (1991) of marked differential

sediment thicknesses of Turonian strata across the

structure. Gardner (1995) identified a number of de-

positional disconformities within lower and middle

Turonian sediments that underlie the upper Ferron

Sandstone Member that he attributed to periodic tec-

tonic uplift. More recently, Martinson et al. (1998)

proposed the existence of subtle bathymetric relief

during Coniacian and Santonian times based on

cross-structure differential subsidence patterns and

detailed foraminiferal analysis. Elsewhere within the

Western Interior, similar intrabasinal structures are

believed to have had a syn-depositional influence on

sedimentary sequences (Merewether and Cobban,

1986; Schwartz and DeCelles, 1988; Heller et al.,

1993). Typically, the age of these culminations pre-

date or approximate to ages of Sevier-style thrusting

(Merewether and Cobban, 1986; Heller and Paola,

1989; Eaton and Nations, 1991). Structural modelling

studies of Heller et al. (1993) and Stephenson and

Cloetingh (1991) suggests that they may be the man-

ifestations of intra-plate contractional stresses that

Fig. 3. High altitude aerial photograph draped on digital terrain model of the Farnham Dome area at the northern tip of the San Rafael Anticline.

The lower Ferron Sandstone is superbly exposed in three dimensions in canyons and cliffs around the periphery of the anticlinal structure. The

core of the structure exposes Early Cretaceous and older rocks. Sections A and B are shown for reference. Vertical scale=5� horizontal.


emanated from the fold-thrust belt and became

uplifted along lines of crustal and lithospheric weak-

ness. These zones of differential uplift are either

equant or elongate features that have broad wave-

lengths (10s of km) and amplitudes ranging from

metres to a few 10s of metres (Merewether and Cob-

ban, 1986; Stephenson and Cloetingh, 1991; Heller et

al., 1993). Ultimately, many of these sites of reported

gentle intraforeland uplift eventually became large,

basement-involved Tertiary dLaramideT structures

(Schwartz and DeCelles, 1988).

2. Sedimentology

The facies associations introduced below are sub-

divided into those belonging to the lower Ferron

Sandstone Member and those of the overlying Juana

Lopez Member (Figs. 4 and 5). The lower Ferron

Sandstone Member comprises (A) Teichichnus-bur-

rowed siltstones which encase, or grade into deposits

consisting of (B) Bioturbated fine-grained sandstones,

(C) Hummocky cross-stratified sandstones, (D) Swa-

ley cross-stratified sandstones, (E) Coarse-grained,

cross-bedded sandstones and (F) Interbedded cur-

rent-rippled sandstones, siltstones and mudstones.

The Juana Lopez Member consists of a single facies

association comprising (G) Organic-rich laminated

shales and interbedded rippled sandstones. The reader

is also referred to Cotter (1975) and Riemersma and

Chan (1991) for sedimentological descriptions and

regional correlations of the lower Ferron Sandstone

around the northern San Rafael Anticline. Bioturba-

tion grade is reported using the scheme of Taylor and

Goldring (1993) where a numeric scale from 0 to 6

(lowest to highest) determines burrow intensity, bur-

row overprinting and the level of preservation of

primary sedimentary structures.

2.1. Lower Ferron sandstone member facies

associations

2.1.1. Facies association A: Teichichnus-burrowed

siltstones

The lower Ferron Sandstone is volumetrically

dominated by grey, massively-bedded, intensely bio-

Fig. 4. Example section through the lower Ferron Sandstone at Section A highlighting the principal facies associations and sequence

stratigraphic elements.


Fig. 5. Key to symbols in Figs. 4, 7, 8, 9, 12, 13 and 14.


turbated, silty mudstone and muddy siltstone and

constitutes the bulk of the recessively-weathered

Tununk Shale. This facies association is pervasive

along the length of the outcrop belt. In more basin-

ward (i.e., eastward) settings the unit thins and the

sharp basal contact becomes increasingly diffuse to

gradational with sandier facies of the lower Ferron

Sandstone. Although dominantly siltstone, the asso-

ciation may coarsen to very fine-grained sandstone

or fine, laminated, dark grey or black claystones.

Thin bentonite ashes form cm-thick, white and or-

ange bands within the grey mudstones. The succes-

sion is easily identified by the relatively high

abundance and low diversity of ichnospecies. The

dominant species is Teichichnus and bioturbation

indices of greater than 4 throughout the succession

produce homogeneous siltstones commonly lacking

primary sedimentary structures and lamination.

Where rare lamination is preserved, faint hummocks

and symmetrical ripple laminations are identified as

subtly coarser bands of coarse silts or very fine

sands against the more commonly encountered

grey fine silts.

Fig. 6. Paralic facies of the lower Ferron Sandstone. (A) Intensely bioturb

sediment deformation features such as folds (arrow) and overlying dew

deposited in an offshore transition zone to lower shoreface setting. (B) T

scale) showing large-scale pillows and dewatering on the eastern flank of

truncated (hashed line) by subsequent flow events. (D) Bioturbated bed at t

interpreted as a surface of submarine sediment bypass containing a high d

(arrow). (E) Thick accumulation of swaley cross-stratified (SCS) sandston

SCS sandstones are underlain by an erosive surface, bearing, large, metre

2.1.2. Interpretation

These bioturbated siltstones were deposited from

suspension in a fully marine, offshore shelf environ-

ment when deltaic systems were confined to western

and northern areas adjacent to the thrust front. Distal

deltaic plumes and/or storm-induced agitation of fine-

grained terrigenous material in proximal areas trans-

ported these sediments basinward in suspension.

Thus, these sediments are interpreted to have accu-

mulated below storm wave base in a low-energy

marine setting. Very distal influences of oscillatory

currents and minor sediment reworking by storm

waves higher in the shoreface profile transported

slightly coarser grained clastic sediment offshore, de-

positing it as thin ripple-laminated layers or fine

hummocky cross-stratified bedforms. Well-circulated

waters encouraged bioturbation, giving rise to a

churned substrate and the ubiquitous obliteration of

bedding surfaces. The Teichichnus trace fossil is com-

monly associated with offshore environments and is

thought to be the deposit feeding trace of a worm-like

organism in unconsolidated substrates (Pemberton et

al., 2001). Its occurrence affirms a distal shelf envi-

ronment interpretation.

2.1.3. Facies association B: bioturbated fine-grained

sandstones

Much of the lower Ferron stratigraphy along the

western and eastern sides of the SRA is composed of

a monotonous succession of well-sorted, fine-

grained, intensely bioturbated sandstone with a fria-

ble weathering character. This homogeneous unit

may reach thicknesses exceeding 8 m that grades

southeastwards (i.e., down-dip) into bioturbated silt-

stones of facies association A and northwestwards

(up-dip) into numerous hummocky cross-stratified

(HCS) sandstones (facies association C) and is

encased by bioturbated siltstones of facies association

A. The basal surface of the association is sharp and

ated sandstones of facies association B. Note the abundance of soft-

atering structures. These sandstones are interpreted to have been

hick, hummocky cross-stratified (HCS) bed (1.5 m Jacob’s staff for

the Farnham Dome. (C) Deformed HCS sandstones with top surface

op of lower shoreface sandstones (facies association C) at Section B,

iversity assemblage of ichnofauna, including large Rosellia burrows

es of facies association D (western flank of the Farnham Dome). (F)

-wide gutters at Section A.



flat above underlying siltstones, whilst the upper

surface is flat to gently undulatory. Dominant ichno-

fauna include Thalassinoides, Ophiomorpha, Palaeo-

phycus, Teichichnus and accessory Arenicolites.

Bedding is poorly developed as a consequence of

ubiquitously high bioturbation indices, commonly

attaining index values of 5 or 6. However, several

faint bedding surfaces are present defined by thin,

cm-thick, discontinuous siltstone interbeds. Where

primary lamination is better preserved, the unit dis-

plays widespread and complex internal deformation

with up to metre-wavelength and decimetre ampli-

tude recumbent folds and abundant water escape

structures (Fig. 6A). Elsewhere, plane-parallel lami-

nation or low amplitude HCS is very rarely and only

partially observed.


This facies association is interpreted to have been

deposited in an offshore transition zone setting beneath

fair-weather wave base. The presence of HCS indi-

cates deposition by episodic storm events and rework-

ing and transport of sediment higher to the distal

reaches of the shoreface profile (Walker and Plint,

1992). Rapid deposition of sands above a substrate

of low strength (most likely siltstones of facies asso-

ciation A) promoted soft-sediment deformation and

rapid dewatering. Variations in bioturbation intensity

in lower shoreface bedforms possibly reflects an in-

verse relationship between storm magnitude and/or

duration between successive storm events and the

degree of biological mixing of the substrate (Dott

and Bourgeois, 1982). Intense bioturbation, reflected

also in the underlying Teichichnus-bioturbated silt-

stones, is likely to have obliterated original bedding

surfaces and internal primary stratification. Homoge-

nization of these beds may therefore be a product of

low magnitude, high frequency storm events, deposit-

ing numerous hummocky cross-stratified sandstone

beds, probably no thicker than a few tens of centi-

metres, with sufficient intervening time for complete

biogenic reworking, in a similar manner to those de-

scribed from the Fulmar Formation of the North Sea

(Howell et al., 1996). Further, high diversity ichnofau-

nal assemblages are regarded as an indicator of water

oxygenation (e.g., Bottjer et al., 1986; Savrda et al.,

1991) that may be promoted by storm surges (Brom-

ley, 1996). The homogenization and relatively thick

accumulation of this association distinguishes these

sandstones from more typical heterolithic lower shore-

face to offshore transition zone strata as predicted by

existing facies models (e.g., Walker and Plint, 1992).

These facies are particularly prevalent along the west-

ern side of the SRA and are characterized by numerous

calcite-cemented botryoidal concretions that were de-

termined by McBride et al. (2003) to have precipitated

in the presence of interstitial brackish water pore

fluids, the significance of which is discussed later.

2.1.5. Facies association C: hummocky cross-strati-

fied sandstones

Sandstones of facies association B grade north-

westwards (i.e., up-dip) into interbedded fine-grained

hummocky cross-stratified (HCS) sandstones and bio-

turbated siltstones. The sandstones are sharp-to ero-

sively-based, light brown or beige-coloured and are

broadly sheet-like occurring as beds that pinch and

swell along their lengths. Bed thicknesses range from

15 cm in the troughs of swales thickening to 60 cm at

the crest of the hummocks. Bioclastic material is

contained within these beds and consists of broken,

disarticulated, thick-walled bivalve and occasional

ammonite debris. Bed tops are sharp and may be

characterized by symmetrical and less commonly

asymmetrical ripple laminae. Internal laminae of the

HCS beds are generally well preserved. Soft-sediment

deformation features are pervasive around the periph-

ery of the Farnham Dome in the north of the study

area. These include large pillows of up to several

metres in width in individual beds up to 1 m thick

(Fig. 6B) and beds with convoluted, overturned lam-

inae. In the latter, the top surface of the bed has been

truncated by subsequent flow events (Fig. 6C). The

basal surfaces are erosive and commonly ornamented

with prominent, cm-amplitude gutters up to 5 cm

long. In these units Thalassinoides, Planolites,

Palaeophycus, Rosselia, Asterosoma and Teichichnus

species predominate, displaying a bioturbation grade

up to 4 with bioturbation intensity increasing towards

the top of the unit. In addition the uppermost burrows

may be filled by a red-weathered (?ferric) cemented

sandy fill (Fig. 6D). In sections where bioturbation is

considerably lower, species diversity is restricted to

sparse vertical Skolithos extending down from the top

surfaces and horizontal Planolites burrows. Typical

bioturbation values for these beds range from 1 to 2.



Hummocky cross-stratified sandstones are indica-

tive of deposition in the lower part of a wave-or storm-

dominated shoreface (Walker and Plint, 1992). Depo-

sition of continuous sheet sands occurs during storms

when large-scale combined oscillatory flows rework

sediment into hummocky bedforms connected by low-

relief swales with mildly erosive bases and low angle

internal laminae (Dott and Bourgeois, 1982; Harms et

al., 1982; Southard et al., 1990). Wave rippled tops of

the sandstones is a result of reworking from waning

oscillatory currents following storm events (Cheel and

Leckie, 1993). Similarly, asymmetrical ripples pro-

duced by bedload transport of grains under uni-direc-

tional currents are frequently reported in storm

deposits (e.g., Walker, 1984; Brenchley, 1985; Leckie

and Krystinick, 1989). Inter-storm periods are repre-

sented by lower energy, suspension deposition of Tei-

chichnus-burrowed siltstones. The upward thickening

of HCS sandstones and the reduction in thickness of

interbedded siltstones reflects basinward shoreface

progradation. The ichnofaunal assemblage affirms a

shallow marine interpretation of these facies and var-

iations in bioturbation intensity may reflect variations

in the duration of successive storm events, where

longer intervals between storm events leads to a greater

intensity of bioturbation (Dott and Bourgeois, 1982).

2.1.7. Facies association D: Swaley cross-stratified

sandstones

Thick sandstones, up to 10 m thick and consisting

of amalgamated swaley cross-stratified sandstones,

are found only at Section A in the northernmost part

of the study area (Figs. 6E). The base of this arenitic

sandbody is highly undulatory and erosive with

metre-wide, sand-filled gutters (Fig. 6F). Swaley

cross-stratification is characterised by concave up-

ward swales, 10 to 50 cm thick, containing sub-par-

allel to slightly divergent internal laminae. Instead of

passing laterally into convex-up hummocks the struc-

ture is truncated by overlying swales, producing a

thick, continuous succession of fine-grained swaley

cross-stratified (SCS) sandstone. A coarse lag of

abundant bivalve and inoceramid fragments, sharks’

teeth, mudstone intraclasts and scattered chert pebbles

line the basal surfaces of the swales. Ichnospecies are

limited to isolated occurrences of Ophiomorpha and

Skolithos and index values are never greater than 1. A

thick succession of these sandstones at Section A

corresponds to the Farnham Unit as described by

Cotter (1975) and the damalgamated hummocky

cross-stratified sandstone faciesT of Riemersma and

Chan (1991). Intense bioturbation by Palaeophycus,

Ophiomorpha, Skolithos, and Thalassinoides in the

upper 3 m of the sandstone has removed most of the

primary stratification leaving behind a homogenised

sandstone cap. Faint concave-up bedding plane sur-

faces within this bioturbated upper part indicate that

these sandstones were probably deposited as SCS

bedforms. Southeastwards (i.e., down-dip) these sand-

stones pass gradationally into sandstones of facies

association C.


Swaley cross-stratification is associated with storm

events (Leckie and Walker, 1982; Hettinger et al.,

1994) and may be a highly amalgamated form of

hummocky cross-stratification whose occurrence is

commonly associated with storm-dominated shore-

faces (Dott and Bourgeois, 1982; Brenchley et al.,

1986; Walker and Plint, 1992). Consequently this

facies association is interpreted to have been deposit-

ed in a proximal lower shoreface to middle shoreface

position (Walker and Plint, 1992) with sands and shell

debris transported basinward from higher on the

shoreface profile. The abundance of shelly debris

and occasional coarse clasts suggests that the coastline

may have been gravelly or rocky. The erosive basal

contact between underlying offshore siltstones and the

overlying SCS sandstones is interpreted as the effects

of wave ravinement of the sea-floor in relation to an

abrupt shallowing of water depth (e.g., Plint, 1988).

The heavily bioturbated upper 4 m of the unit reveals

intense biogenic reworking by shallow marine ichnos-

pecies towards the end of deposition of this associa-

tion and implies a depositional hiatus towards the top

of the succession.

2.1.9. Facies association E: coarse-grained, cross-

bedded sandstones

Encased within Mancos Shale are coarse-to gran-

ular-grained, poorly to moderately well-sorted, peb-

ble-bearing arenitic sandstones (Fig. 7A). These

sandstones are best exposed west of the Sphinx rail-

way siding (Sections K, L and M) in the southern part

of the study area (Fig. 7B). Mapping of the associa-



tion with integrated palaeocurrent information reveals

a single, low-relief linear body up to 4 km wide

encased within the Mancos Shale, the southern edge

of which is exposed at Section N (Fig. 7C). The lateral

margins of this gently lensing unit show no interdig-

itation with other facies associations indicating gentle

incision into underlying siltstones of facies association

A (Fig. 7C). The basal surface of the unit is flat to

erosional and the uppermost surface is an equally

sharp contact with open marine mudstones. The unit

comprises irregular beds 1 to 40 cm thick that lens,

split and amalgamate and are intercalated with lesser

thicknesses (up to 5 cm) of Teichichnus-bioturbated

siltstone (Fig. 7D).

The association has been removed in much of the

middle part of the study area due to present-day

erosion but reappears in the north where it consists

of thin, highly amalgamated and erosively-based beds

that reach a cumulative thickness of less than 10 cm

(Fig. 7E). In these localities 15 mm-long sharks’ teeth,

mudstone intraclasts, an exotic array of rounded cher-

tiferous, feldspathic and quartzitic granules, pebbles

and cobble-sized clasts show cross-stratification with

southeastward-dipping foresets. Despite the exotic

collection of pebbles present, these clasts are textur-

ally mature in roundness and sphericity. The unit

bears abundant bioclastic material including fragmen-

ted, disarticulated inoceramid shell debris and large

shark teeth in some instances exceeding 1 cm in

length. Bioturbation is generally absent but rare Pla-

nolites and large Thalassinoides networks occur at the

tops of sandstone bedding surfaces. Mean grain size

decreases southwards from granular, locally conglom-

eratic, pebbly sandstone in the northern outcrops

(Sections C, D and E), to coarse-grained sandstone

at Sections K to M. Sorting also improves southwards

from very poorly sorted in the north to moderately and

poorly sorted in the south.

In Section L the association thickens to a maxi-

mum thickness of 2 m. Here individual beds have

Fig. 7. Facies association E of the lower Ferron Sandstone. (A) Interpreted

encased by offshore siltstones of the Mancos Shale and forms a gently-lens

single 1-cm thick bed at Section N and shows no interdigitation with ot

splitting and amalgamation and contain many erosive surfaces (lines) to pr

sandstones are coarser-grained and poorly sorted (Section C) and show

characterised by erosively-based, ungraded, cross-stratified beds bearing co

sustained turbidity flows. (G) Bed amalgamation resulting from successive

bedded lower part beneath line, plane-parallel laminated upper part).

erosive basal surfaces (Fig. 7F), ornamented by mod-

erately intense tool markings. These structures include

spatulate flutes, discontinuous linear grooves and

broad u-shaped gutter casts, commonly exceeding 5

cm in width and revealing southward palaeoflow

directions. Beds exhibit high degrees of scour into

underlying sandstones, forming locally thickened,

amalgamated packages (Fig. 7D). The sandstone

beds form partially-amalgamated lenses that can be

traced over distances of up to 5 m to many tens of

metres. Cross-stratification is common, but some thin-

ner beds can be massive, plane-parallel laminated or

ripple cross-laminated. Highly irregular internal scour

surfaces are common, producing amalgamated sandy

packages of differing juxtaposed stratification (e.g.,

Fig. 7G). South-to southeastward-dipping centimetre-

wide, sand-rich foresets are defined by thin medium to

fine coarse-grained sandstone laminae that dip at

angle of repose and have tangential to asymptotic

bases indicating both planar and trough-cross stratifi-

cation. Coarser material and clasts of up to 7 cm in

diameter can be found at the toes of foresets. Al-

though some beds show gentle normal grading from

very coarse to coarse-grained sandstones, ungraded

beds are most dominant. In nearly all cases, the top

surfaces of the sandstones are always sharp with

overlying burrowed siltstones.


Erosional bases, dominance of tractional structures

and marine ichnofacies reflect deposition by numer-

ous turbidity flows in an offshore to outer shelf set-

ting, forming coarse-grained barforms. These deposits

have many parallels with the cross-stratified sands of

the turbidite classification scheme of Pickering et al.

(1986) such as mean grain size, sedimentary structures

and the unusually coarse grain size relative to bed

thickness. Scoured bases to the sandstone beds reveal

the erosional capacity of a waxing head of a turbulent

cloud prior to the aggrading, traction-dominated de-

graphic log from Section L near the Sphinx siding. (B) This unit is

ing sandbody over 4 km. (C) The lateral margin of this unit thins to a

her facies associations. (D) Individual beds show high degrees of

oduce thickened, sandy packages. (E) In proximal areas, equivalent

faint cross-bedding. (F) In more distal areas, the association is

bbles (encircled) and sharp tops and are interpreted as the products of

erosive flows juxtaposes contrasting sedimentary structures (cross-


positional phase. The assimilation of sharks’ teeth into

these deposits and the high degrees of scour and

amalgamation in the northern part of the study area

imply cannibalisation of the sea-bed by an erosive

flow prior to deposition (i.e., the flows were

bignitiveQ, sensu Parker, 1982). Up-dip thinning of

the unit is interpreted to reflect greater degrees of

submarine erosion and bypass as a function of greater

flow velocities.

Bedload deposition behind the head of the turbu-

lent cloud records the downstream migration of

straight to sinuous-crested dune forms producing

cross-bedding, and coupled with the absence of either

normal or reverse grading suggests that the flows were

sustained rather than unsteady and waning (sensu

Kneller, 1995; Kneller and Branney, 1995). Gentle

sediment grade changes with down-dip distance

have also been attributed to the effects of flow conti-

nuity (Plink-Bjorklund and Steel, 2004). The presence

of outsized cobble clasts lying many tens of km into

the basin indicates that the velocities of the flows were

also very high (at least 1.7 m s�1 given the modal

grain size of the sandstones and the sedimentary

structures present) and continuous over long dis-

tances. In contrast, there is little evidence for deposi-

tion by waning flows, such as composite upward

successions of progressively finer grain sizes and

lesser stability bedforms or Bouma sequences. How-

ever, interflow periods are represented by passive

suspension settling of thin siltstones that separate the

sandstones. These may become wholly or partially

removed during subsequent flow events.

Development of dune-scale bedforms is a relatively

unusual phenomenon in turbidite systems. Very large-

scale bedforms with wavelengths of tens to hundreds

of metres have been identified on the surfaces of

modern lower slopes and turbidite fans (Prior et al.,

1986; Bornhold and Prior, 1990; Hughes-Clarke et al.,

1990). It is likely that such forms require relatively

continuous, high velocity flows. Channels are an im-

portant morphological element in turbidite systems as

they contribute to enhancing the long-distance trans-

port capacity of coarse sands and pebbles by maximiz-

ing flow velocity (Postma et al., 1988). The lenticular

cross-sectional geometry of this facies association in

the vicinity of the Sphinx siding suggests that deposi-

tion occurred within a channel or chute. However, the

broad, shallow dimensions of this association is some-

what unusual and probably reflects the natural shal-

lowing and widening of submarine channels with

downslope distance (e.g., Clark and Pickering,

1996). At down-dip channel termini, turbidity flows

begin to expand and become net depositional. Exper-

imental analogues have indeed shown that at these

points of expansion, both sinuous and straight crested

bedforms may readily develop (Imran et al., 2002).

Although turbidity currents are known to be capa-

ble of transporting clasts of pebble size and greater,

resting positions at the toes of foresets suggest bed-

load transport. Coarse grain sizes can be maintained in

suspension by turbulence, buoyancy and hindered

setting support mechanisms (Pickering et al., 1986).

Clasts are not found bfloatingQ within beds, implying

that these facies were not deposited by non-turbulent,

laminar inertia-flow (or traction carpets) as simulated

experimentally (Postma et al., 1988). The position of

outsized clasts at the toes of foresets, together with an

absence of other distinguishing features of laminar,

plastic rheology flows (i.e., poor sorting, random

clastic alignment, massive beds displaying reverse

grading, plastic deformation features and so on, e.g.,

Major, 1997; Shanmugam et al., 1995; Shanmugam,

1996) indicates bedload-dominated deposition at the

base of a fluidal, turbulent flow of considerable dura-

tion rather than single pulse-like flow events. Further,

flows that are underladen with respect to their carry-

ing capacity, result in a predominance of tractional

process at the base of the flow and the production of

dune-scale bedforms (Pickering et al., 1986).

Hiscott et al. (1997) listed several criteria to ac-

count for the origin of sharp-topped turbidites, such as

the abrupt termination of deposition of an aggrading

bed, bimodal grain size distributions or longitudinal

grain size partitioning between the head and tail of the

flow. In addition, sharp bed tops might also reflect

flow-lofting processes when deposition at the base of

the flow causes the relatively low density interstitial

fluid to convect upwards, producing a buoyant plume

that lofts off the upper surface of the flow and result-

ing in the enhancement of grain size contrasts between

turbidite sandstones and overlying finer-grained units

(Sparks et al., 1993). Much of the finer-grained frac-

tion carried by the plume may then become dispersed

over wide areas (McLeod et al., 1999). Finally, abrupt

grain size contrasts may reflect the removal of finer-

grained material by flow-stripping over channel mar-


gins, resulting in the enhancement of the lower, sand-

rich part of the flow (Piper and Normark, 1983;

Bowen et al., 1984). However, it should be stressed

that channel sinuosity has not been directly observed.

Consequently, whilst flow stripping is an integral part

of submarine channel evolution and fan growth (Hay,

1987), here the process is only inferred.

Riemersma and Chan (1991) and Molenaar and

Cobban (1991) proposed that this coarse sandy inter-

val represent a transgressive lag resulting from a rise

in relative sea-level. However there is no evidence for

lowered wave base and wave impingement on the sea-

floor prior to their deposition and the process of

winnowing subjacent deposits (siltstones) during

transgression would not produce the grain sizes ob-

served. In addition, terrestrial (e.g., single or multi-

storey fluvial sandstones, interfluvial palaeosols and

rooted horizons) and estuarine (e.g., clay-draped tidal

sandstones, bi-directional current indicators, bay and

lagoon mudstones) deposits are absent in these rocks,

precluding their interpretation as the products of val-

ley incision and filling.

2.1.11. Facies association F: interbedded current-rip-

pled sandstones, siltstones and mudstones

This facies association is found at Sections F

through J and again at O. Spatially, it can be found

laterally adjacent to the coarse-grained planar cross-

bedded sandstones. It consists of a 1 to 1.5 m thick

heterolithic succession comprising parallel-and cur-

rent-ripple laminated, very fine-to medium-grained,

normally graded sandstones interbedded with biotur-

bated siltstones (Fig. 8A and B). Sandstone beds

occupy 20 to 50% of the facies association. In sand

poor areas, such as at Section O (Green River) the unit

is populated by numerous concretions of decimetre

diameter. This association also thins northwards, but

rather than becoming coarser grained to the north, the

sandstones maintain a near-uniform grain size across

the entire outcrop belt.

Individual sandstone beds typically range in thick-

ness from 1 to 4 cm although rare larger beds reach

thicknesses of up to 9 cm. Bed geometries vary in

equal measure from discontinuous and lenticular, to

continuous and sheet-like, but bed amalgamation is

rare. Lenticular beds may form small, decimetre-long

ripple-laminated sandstones isolated in bioturbated

siltstones (Fig. 8C). In all cases, bed basal surfaces

are characterised by flat or gently scoured bases with

abundant and varied arrangement of grooves, flutes

and prod marks and indicate southerly flow directions.

These tool marks have mm-scale amplitudes and

widths and axial lengths of less than 5 cm. Typically

beds are ripple laminated, however the thicker beds

may consist entirely of plane-parallel laminations

(Fig. 8D) or upward graded parallel-to ripple-lamina-

tion. Asymmetric ripples are no more than 2 cm in

amplitude and greater than 10 cm in wavelength and

display south-to southeasterly-dipping foresets. Stoss-

side preservation is rare. Mean palaeocurrent values

show a 268 divergence between sole mark and ripple

bedform measurements (Fig 8E). Beds are normally

graded or ungraded. Where grading is observed, the

upward transition from sandstones into overlying silt-

stones is abrupt. In ungraded beds, the top surface of

the bed maintains asymmetrical current ripple bed-

form morphology and has a sharp contact with over-

lying siltstones. Trace fossils are generally rare within

the sandstones, but isolated examples of vertical bur-

rows (possibly Diplocraterion) and small crawling

traces are present.


Scoured bases, normal grading, and composite ar-

rangement of parallel and ripple-laminated sandstones

are all suggestive of transport and deposition by wan-

ing turbidity flows. These facies are comparable to thin

bedded sand–mud couplets facies in the classification

of Pickering et al. (1986). Episodic southward flows

appear to have transported fine-grained sand over large

areas. Planar-and ripple-laminated sandstones there-

fore correspond to Tb and Tc divisions of partial

Bouma sequences (Bouma, 1962) and indicate bed-

load sediment transport. Tbc sequences indicate that

these flows were capable of producing both upper flow

regime plane bedding and sub-critical lower flow re-

gime ripple-drift bedforms. The common presence of

lenticular bedding suggests fluctuations between sed-

iment supply and flow strength, most likely indicating

that flows were underladen with respect to their carry-

ing capacity. Tool marks at the bases of sandstone beds

exhibit the latest, pre-depositional scouring effects of

turbulent uni-directional flows prior to current-ripple

settling and migration from the base of the flow. The

variance in mean palaeocurrent directions taken from

these structures compared to palaeocurrents taken

Fig. 8. Facies association F of the lower Ferron Sandstone. (A) Interpreted graphic log from Section O near Green River. (B) Heterolithic

succession consisting of graded sandstone beds. Sandstones are erosively-based, plane-parallel and ripple laminated and are interpreted as the

products of waning turbidity flows. (C) Beds may be thin, lenticular and ripple-laminated or (D) thicker, more sheet-like and plane-parallel

laminated. (E) Palaeocurrent data show flows were towards the south and southeast. Sole structures show a southerly flow direction, whereas

palaeocurrents from cross-strata show flow directions towards the southeast (incorporates palaeocurrent data from facies association E; mean

values provided).


from ripples indicates subtly different orientation of

incident erosive flows from the subsequent deposition-

al flow. Mudstone interbeds probably reflect deposi-

tion out of suspension of fine-grained material

entrained within the tail of the flow. Since this facies

association can be found adjacent to facies association

E, it is considered to be the result of unconfined over-

bank or interchannel deposition. Collectively the two

associations constitute genetically-related depositional

elements (after Mutti and Normark, 1991) of a chan-

nel-overbank system.

These facies also have many indicative features of

tempestites. Despite their common association with

combined or oscillatory currents, storm-generated


tempestite deposits may comprise turbidite-like beds

consisting of tool-marked bases and asymmetrical

ripples (Hamblin and Walker, 1979) with long, high

velocity run-out distances in excess of 100 km (Walk-

er, 1984; Brenchley, 1985; Leckie and Krystinick,

1989; Myrow and Southard, 1996) often parallel to

the shoreline trend (Walker and Plint, 1992). To this

extent, facies association F appears similar to current-

modified tempestites (Myrow and Southard, 1996).

However, the normal-to-shoreline palaeoflows and

synchronous deposition of laterally adjacent coarse-

grained sandstones are not compatible with deposition

by sheet-like, storm-generated currents.

2.2. Juana Lopez member facies association

2.2.1. Facies association G: organic-rich laminated

shales and interbedded rippled sandstones

Dark grey to black, clay-rich mudstones dominate

this facies association and are characterized by an

absence of bioturbation, making it distinct from

deposits of the underlying lower Ferron Sandstone.

Hence the basal surface of the association is distinc-

tive where it forms a textural contrast with underlying

sandstones and Teichichnus-bioturbated siltstones be-

longing to the lower Ferron Sandstone and is easily

identified. The finely laminated shale is extremely

fissile, contains rare, centimetre-thick sandstone

lenses and dominates the lower half of the member

(Figs. 9 and 10A). Southward thickening is observed

in the lower part of the interval from 3 m in north-

ernmost sections to 6 m in the vicinity of Green River.

The upper half of the member also comprises

fissile black shales but is interbedded with numerous

very-fine and fine-grained sandstones (Figs. 9 and

10A). Towards the top of this heterolithic succession,

the association weathers to a lighter brown colour and

becomes increasingly fossiliferous with abundant and

well-preserved ammonite, bivalve and gastropod de-

bris distributed along bedding surfaces. Individual

sandstone beds are typically sheet-like, 1 to 4 cm

thick and separated by shale interbeds of centimetre-

scale thicknesses. The internal fabric of the sandstones

is similar to those of facies association F, comprising

both current ripple-and plane-parallel laminated sand-

stones. Parallel-laminated beds are occasionally found

to grade upwards into current-ripple cross-stratifica-

tion and undulatory ripple-tops. Many of the ripple-

laminated beds occur as discontinuous decimetre-long

lenses. Bed bases are intensely scoured with prod,

groove and flute markings, providing reliable palaeo-

flow information (Fig. 10B). The geometries of the

sole marks in this succession display amplitudes and

widths of mm scale and wavelengths that are limited

to a few cm. Tool mark and ripple lamination palaeo-

current indicators repeatedly show discordance in ori-

entation values within individual bed readings. The

mean palaeoflow orientation attained from sole struc-

tures is 1728S compared to ripple lamination measure-

ments indicating south–southeast flows, averaging

1578S (Fig. 10C). Repetitive sandstone–shale couplets

occur as concentrated packages up to 3 m thick,

separated by papery shale horizons less than 50 cm

thick with widely distributed centimetre-thick, very

fine-grained, ripple-laminated sandstone beds (Fig.

10A). These packages may show coarsening-and-

thickening upward bedset trends, in addition to fin-

ing-and-thinning upward series of bedsets. Packages

become progressively sandier and thicker upwards

although much of the top of the Juana Lopez Member

is poorly exposed due to removal by recent erosion.


These facies are remarkably similar to those of

facies association F and thin bedded sand–mud cou-

plets of Pickering et al. (1986). However their attri-

butes are also remarkably similar to documented

examples of tempestites (Myrow and Southard,

1996). As such this association could be interpreted

either as turbidites or storm-related tempestites for

which an understanding of relative sea-level context

is of critical importance. Nonetheless, the recurring

nature of sand depositional events are interpreted as

the products of repetitive initial erosive flows that

subsequently wane and become depositional. Sole

structures reveal the scouring effects of southward-

directed flows at the waxing head of an erosive tur-

bulent cloud. The sandstone beds record the waning

of uni-directional flows, represented in the transition

from horizontally-laminated to ripple-laminated sand-

stones. These flows appear to have been dilute and

sediment starved as observed by the discontinuous,

ripple-laminated, sandstone lenses. The accumulations

of laminated claystones represent periods of pelagic

deposition between flows with little contamination

from clastic sediment input.

Fig. 9. Example measured section through the Juana Lopez Member at Section N illustrating the principal facies and their interpretations.


Outflow distances extend many tens of km south

and eastwards beyond the study area, attesting to

widespread, sheet-like sandstone deposition. Similar

to the observed disparity of orientation of directional

data in the lower Ferron deposits, a 158 mean differ-

ence between tool mark and ripple lamination palaeo-

currents reflects subtly differing orientations of the

initial erosive flow and subsequent deposition of

deflected bedload as flow energy dissipated. The bun-

dles of couplets representing numerous flow events are

interpreted as parasequences, punctuated by short-

lived recourses to mudstone deposition, interpreted

as flooding surfaces and parasequence boundaries.

2.3. Lower Ferron sandstone stratigraphic architec-

ture and chronostratigraphy

The sequence architecture of the lower Ferron and

Juana Lopez stratigraphy is summarised below and

illustrated in Figs. 11, 12, 13, 14 and 15. In northern

Castle Valley the lower Ferron Sandstone consists of

the shallowest water facies, namely facies associations

B, C and D (Fig. 11A) which thin southeastwards

until they eventually grade into marine siltstones of

facies association A (Fig. 11B). Southeastwards, these

facies grade into turbidite deposits of facies associa-

tions E (Fig. 11C) and F (Fig. 11D). The lower

Fig. 10. Facies of the Juana Lopez Member. (A) The heterolithic interval consists of numerous sandstone-mudstone bundles separated by mud-

rich sections and is interpreted as a parasequence. Individual sheet-like and lenticular beds are graded and plane-parallel to ripple laminated.

Parasequences thicken and coarsen upward through the succession. (B) The basal surfaces of beds are intensely scoured with abundant sole

structures. (C) Palaeocurrent data from sole marks show southerly flow directions whilst cross strata show flows to the southeast (mean values

provided).


Ferron–Juana Lopez succession comprises one com-

plete sequence broken down into component systems

tracts. Initial highstand, falling stage, lowstand and

transgressive systems tracts define the lower Ferron

Sandstone whilst only a highstand systems tract repre-

sents the Juana Lopez Member (Fig. 11A–D).

2.3.1. Highstand systems tract

The lower Ferron shoreface deposits are character-

ized by both coarsening upward intervals of inter-

bedded HCS sandstones and burrowed siltstones

(facies associations A and B) and erosively-based

shorefaces consisting of amalgamated SCS sandstones

(facies association D). Normal progradation is repre-

sented by upward coarsening interbedded HCS sand-

stones and bioturbated siltstones that record the

systematic outbuilding of the shoreline as sediment

flux exceeded accommodation creation (Walker and

Plint, 1992). Termination of the coarsening upward

trend, and replacement by thicknesses of offshore

bioturbated siltstones indicates a rise in relative sea-

level and therefore defines a flooding surface. Each

cycle is interpreted as a storm-dominated parase-

quence (Van Wagoner et al., 1990). The vertical stack-

ing of successively thicker parasequences indicates

net progradation of the shoreline and defines the distal

portion of a highstand systems tract (Van Wagoner et

al., 1990) at the base of the lower Ferron succession

(Fig. 11A and Section A of Fig. 12).

2.3.2. Falling stage systems tract

Sharp-based middle shoreface SCS sandstones

scour into underlying offshore siltstones, and in their

most proximal reaches omit lower shoreface intervals

predicted by normal progradational facies models

(e.g., Walker and Plint, 1992). When traced palaeo-

Fig. 11. Outcrop expression of key stratigraphic surfaces. (A) Upward-thickening and coarsening sandstones of facies association C at Section

A define highstand systems tract parasequences bounded by flooding surfaces. The top of the cliff comprises a thick stack of sandstones of

facies association D. The basal erosive surface and abrupt juxtaposition of facies association D on facies association C is interpreted as a

regressive surface of marine erosion resulting from a lowering of relative sea-level. The overlying shallow marine sandstones are thus

interpreted as the products of a falling stage systems tract. (B) These sandstones grade down-dip into siltstones of the Mancos Shale. Overlying

highstand systems tract deposits of the net-progradational Juana Lopez Member are clearly seen in the top of the cliff. A sequence boundary,

characterised by submarine bypass and erosion, separates the falling stage and highstand systems tract. Photo is taken near Section D. (C)

Down-dip correlative strata to the sequence boundary thicken and comprise turbiditic sandstones of facies association E and (D) laterally

adjacent facies association F.


seaward, the amalgamated SCS sandstones grade into

HCS sandstones of facies association C (Figs. 11A

and B and Section B, Fig. 12). Shoreface sandstones

belonging to this genetic unit are represented over

several km further basinward than deposits of the

highstand systems tract. The long-distance regression,

omission of lower shoreface facies, erosive base

caused by wave ravinement and down-dip translation

Fig. 12. Stratigraphic correlation panel of the lower Ferron Sandstone in t

represented by deposition of highstand and falling stage systems tract

shoreface units and separate falling stage shoreface deposits form a thin s

transgressive systems tract strata are thus interpreted to downlap onto old

and are interpreted to onlap older shoreface strata in a landward direction.

the Farnham Dome structure (Section B). (B) Map provided for reference

into an apparently conformable succession of shore-

face strata defines an abrupt shallowing of water depth

associated with forced regression (Plint, 1988). Con-

sequently the basal surface is interpreted as a regres-

sive surface of marine erosion and the overlying

sandstones constitute a falling stage systems tract

(Plint, 1988; Hunt and Tucker, 1992; Posamentier et

al., 1992; Plint and Nummedal, 2000).

he northern part of the study area. (A) Initial shoreline regression is

strata. Deposition of lowstand turbidites overlies these regressive

uccession of transgressive systems tract deposits at Section A. The

er deposits. The lowstand turbidites thicken in a down-dip direction

Thinning of falling stage systems tract strata can be observed across

.


Fig. 13. Stratigraphic correlation panel of the lower Ferron Sandstone in the southern part of the study area. (A) Lowstand systems tract strata

comprising facies associations E and F thicken basinwards. The lenticular distribution of facies association E and the absence of interdigitation

with facies association F suggest the presence of a broad, low-relief channel. Collectively the two associations constitute genetically-related

depositional elements of a channel-overbank system. (B) Map provided for reference.


2.3.3. Lowstand systems tract

Sandstones of facies association E can be traced

up-dip where they comprise a thin gravelly cap above

falling stage bioturbated sandstones (Sections C, D

and E, Fig. 12). Because of this relationship, the

turbidites of facies associations E and F are not con-

sidered contemporaneous with deposition of the un-

derlying forced regressive shoreface (Fig. 12). In their

justification of a fourfold systems tract model, Plint

and Nummedal (2000) argued that turbidites lying

basinward of forced regressive shorefaces constitute

lowstand systems tract strata since their deposition

post-dates forced regression. This criterion is adopted

here, and the lower Ferron turbidites are thus consid-

ered to represent a period of lowstand deposition

beyond the basinward limits of the underlying falling

stage strata (Figs. 13 and 15). Consequently the fall-

ing stage and lowstand systems tracts are separated by

an interpreted sequence boundary. In up-dip sectors

this sequence boundary is manifest as a submarine

bypass surface mantled by a thin gravel laggy deposit

of facies association E and intense bioturbation into

underlying falling stage shoreface sandstones (Sec-

tions A to E, Fig. 12). We stress that other features

associated with relative sea-level lowstands (incised

valleys, well-drained palaeosol horizons) are not iden-

tified, indicating that the lowstand shoreline did not

extend basinwards as far as the present day outcrop. A

downslope increase in the thickness of the channel–

overbank depositional elements may be attributed to a

gradient change, analogous to a transition from the

base of slope to the relatively flat basin floor and the

downslope transition from dominantly erosional to

dominantly depositional regimes.

2.3.4. Transgressive systems tract

The top of the lower Ferron succession consists of

a thin heterolithic succession comprising HCS sand-

stone and bioturbated siltstone interbeds. These facies

sit atop bioturbated sandstones capping the forced

regressive strata (Section A, Fig. 12). The return to

lower shoreface deposition at Section A denotes a rise

in relative sea-level and the presence of a flooding

surface at the base of a distal storm-dominated para-

sequence. The flooding surface merges with the un-

derlying sequence boundary and separates falling

stage deposits from the overlying transgressive sys-

tems tract deposits. Sand-poor, laminated mudstones

of the Juana Lopez that overlie the lower shoreface

interval, indicate a landward shift of the shoreline in

Fig. 14. Stratigraphic correlation panel of the Juana Lopez Member along the length of the study area. Progradational stacking of parasequences

indicates deposition under highstand conditions following transgression of the lower Ferron shoreline. (B) Map provided for reference.


response to continued sea-level rise and are inter-

preted as the maximum flooding surface of the

lower Ferron succession.

2.4. Juana Lopez member stratigraphic architecture

and chronostratigraphy

2.4.1. Highstand systems tract

Deposition of Juana Lopez turbidites that overlie

the lower Ferron Sandstone can be separated into two,

upper and lower architectural divisions (Fig. 14). The

lower part of the succession is represented by contin-

ued mud-dominated deposition immediately follow-

ing the transgression at the end of lower Ferron

Sandstone time, with sandstones of patchy distribu-

tion confined to central and southern sectors of the

study area. The upper part of the Juana Lopez suc-

cession is characterized by northwest to southeast

progradational stacking of parasequences. This sea-

ward stepping architecture suggests basinward progra-

dation as sediment supply outpaced accommodation

creation associated with conditions of sea-level high-

Fig. 15. Chronostratigraphic diagram summarising the spatial and temporal changes in deposition of the lower Ferron and Juana Lopez

Members. Importantly, the deposition of facies associations E and F that constitute the lowstand systems tract of the lower Ferron Sandstone

correspond to a surface of non-deposition and bypass in up-dip areas. The absence of co-genetic slumped deposits suggests that a direct

connection between lowstand rivers and the deeper basin existed. Consequently these turbidites are interpreted to have been initiated by

hyperpycnal flows exiting flooded river mouths.


stand. Regional correlations reveal that on the western

side of the SRA the Juana Lopez Member corresponds

to outer shelf and turbidite-free mudstones of the

Washboard Unit in northern Castle Valley (Molenaar

and Cobban, 1991; Riemersma and Chan, 1991; Gard-

ner, 1995) and pronounced seaward stepping of the

upper Ferron delta in southern Castle Valley (Gardner,

1995). This seaward stepping was associated with

paralic and coastal plain accretion and stratigraphic

climb (Gardner, 1995), indicating the progradation of

the shoreline was coupled with rising sea-level typical

of highstand conditions (Van Wagoner et al., 1988;

Van Wagoner et al., 1990).

3. Discussion

The interpretation of sandstones deposited by tur-

bulent flows demands further discussion regarding the

propensity of shallow marine systems to deliver sand-

grade sediment via turbidity currents to offshore

environments over long distances during forced re-

gression, despite the absence of a clearly discernable

shelf break. Any depositional model for the lower

Ferron Sandstone needs to account for the initiation

mechanism for the generation of turbidity currents

during falling and lowstand of relative sea-level in a

basin lacking a shelf edge. Additionally, the signifi-

cance of lower Ferron and Juana Lopez shelf sand-

stone distribution in central Utah, and the possible

cause of variance in palaeocurrent directions between

sole marks and cross-stratification, are discussed.

3.1. Turbidite initiation mechanisms

It is widely regarded that turbidites form as a

consequence of either delta front failure and mass

gravitational sliding and slumping, or by hyperpycnal

underflows resulting from high sediment loads

debouching at flooded river mouths (Normark and

Piper, 1991). The type of flow generated by these

two separate mechanisms, and the types of bedforms

and facies that they produce, will determine flow

continuity and steadiness (Kneller, 1995; Kneller

and Branney, 1995). It is common belief that turbidity

currents are initiated when sea-level falls to, or be-

yond the continental shelf edge (Mutti, 1985; Mutti

and Normark, 1991; Posamentier et al., 1991; Weimer,

1991; Normark et al., 1993). Slides and slumps result-

ing either from earthquake seismicity or depositional

oversteepening of the delta front at the top of the

continental slope transform into debris flows, and

ultimately, surge-type turbidity flows of several

hours duration (e.g., Piper et al., 1988; Normark and

Piper, 1991; Mulder and Syvitski, 1995). Surge-type

flows are typically depletive as current mixing with


ambient water results in a loss of density contrast,

sediment concentration and velocity and will produce

a predictable arrangement of facies that records initial

seabed erosion by the waxing head of the turbulent

cloud, followed by deposition of normally graded

sequences in the depletive tail of the flow (e.g.,

Bouma, 1962; Lowe, 1982). The absence of regional

shelf edge physiography inhibits the development of

failure-related turbidites in ramp margins such as the

Western Interior Basin, and as such, turbidites are

relatively rare.

Kneller (1995) and Kneller and Branney (1995)

showed that depletive, unsteady flows typical of

surge-type turbidity currents represent one flow struc-

ture scenario from at least 13 different possible flow

types. Furthermore, a number of recent studies have

suggested that turbidity currents may last for consid-

erably longer durations than the traditionally-held

waning-flow perception (e.g., Kneller, 1995; Kneller

and Branney, 1995; Mulder and Syvitski, 1995;

Mulder et al., 1998; Kneller and Buckee, 2000;

Mulder and Alexander, 2001; Parsons et al., 2001;

Mulder et al., 2003 and many others). This conclusion

is drawn from a wide range of approaches including:

(1) studies of highly sinuous turbidite channels

(Damuth et al., 1988), (2) direct observations and

sampling from lakes, high-latitude fjords and marine

basins (Weirich, 1986; Hay, 1987; Wright et al., 1988;

Prior and Bornhold, 1989; Chikita, 1990; Prior and

Bornhold, 1990; Zeng et al., 1991; Piper and Savoye,

1993; Piper et al., 1999; Kineke et al., 2000; Hicks et

al., 2004), (3) laboratory experiments (Rimoldi et al.,

1996; McLeod et al., 1999; Alexander and Mulder,

2002), (4) depositional and physical stratigraphic rela-

tionships (Plink-Bjorklund and Steel, 2004), (5) bed-

forms not compatible with waning flow processes

(Kneller, 1995; Kneller and Branney, 1995; Nemec,

1995; Mulder et al., 1998; Piper et al., 1999; Mellere

et al., 2002; Mutti et al., 2003; Plink-Bjorklund and

Steel, 2004) and (6) numerical models (Chao, 1998;

Imran et al., 1998; Kassem and Imran, 2001). The

maintenance of sustained flows is most commonly

attributed to the development of hyperpycnal under-

flows at river mouths, that may last for many hours,

days or even weeks (Mulder et al., 1998).

Hyperpycnal flows initiate when low salinity river

discharge enters marine waters with suspended sedi-

ment concentrations in excess of 36 kg m�3 (Mulder

and Syvitski, 1995). The excess density enables sed-

iment-laden flows to become negatively buoyant rel-

ative to the ambient, resulting in a long-lasting

seaward flow. From present day observations, these

conditions may be met following major river floods

carrying large volumes of sediment on a periodic

frequency ranging from b1 to N10 years (e.g., Mulder

and Syvitski, 1995; Johnson et al., 2001; Mulder et

al., 2001; Hicks et al., 2004). Some authors have

proposed that hyperpycnal flows are most common

during periods of sea-level highstand (Mulder and

Alexander, 2001) whilst others have argued that

they may be more frequent during sea-level falls

when large parts of the shelf become excavated

(e.g., Normark and Piper, 1991; Plink-Bjorklund and

Steel, 2004). Regardless, the relatively frequent oc-

currence at which these flows occur during present

day sea-level highstand suggests that ancient hyper-

pycnal flows ought to be more common than presently

documented.

3.1.1. Lower ferron sandstone depositional model

The following observations favour a hyperpycnal

flow origin for the lower Ferron turbidites: (1) the

presence of cross-bedded sandstones, (2) paucity of

genetically related slump and debris flow and delta

front/prodelta deposits, (3) surface of sediment bypass

in up-dip sectors, (4) change in dominant shoreline

depositional process, (5) flow channelisation and later-

al partitioning of sustained and waning flow conditions

and (6) regional palaeogeographic considerations.

3.1.1.1. Dune-scale cross-bedding. The relatively

recent recognition of facies pertaining to deposition

by riverine underflows has thus far revealed only a

limited array of types and arrangement of facies. The

most common of these are thick, ungraded sandstone

bodies that may be massive (e.g., Kneller, 1995;

Kneller and Branney, 1995) or plane-parallel laminat-

ed (e.g., Plink-Bjorklund and Steel, 2004). Mulder et

al. (1998) documented a hyperpycnite from the

Saguenay fjord, Canada that was characterised by a

graded motif from coarsening-upward at the base to

fining-upward towards the top. Waxing and waning of

current strength that gave rise to the coarsening up-

ward-fining upward succession was attributed to in-

creasing and decreasing river outflow during the

course of a flood. Plink-Bjorklund and Steel (2004)


suggested that hyperpycnal flow beds may be signif-

icantly more complex, and place emphasis on the up-

dip to down-dip variability of grain size trends and

arrangement of sedimentary structures that deviate

from the Saguenay model.

Mulder and Alexander (2001) proposed that climb-

ing ripples might readily form during hyperpycnal

flow events as a result of sustained flow durations

and sediment transport. Little comment has been made

regarding the propensity of these flows to deposit

larger bedforms such as dunes. However, Mulder

and Alexander (2001) suggested that the production

of dunes is likely to be relatively uncommon owing to

the longer durations required for their development.

The cross-bedded lower Ferron examples may howev-

er, represent an unusual example of dune-scale bed-

form development during the passage of turbidity

currents. Such prolonged bedload sediment transport

is consistent with the longer flow timescales associated

with hyperpycnal flows. However, because dune-scale

bedforms may also be associated with surge-type

flows, additional distinguishing criteria are required

to support a hyperpycnal flow interpretation.

3.1.1.2. Absence of slump, debris flow and delta front/

prodelta deposits. Turbidites deposited as a result of

slope failure are commonly associated with surge-type

flows. Under these circumstances, slumped and de-

formed sandstones ought to be found in association

with equivalent, down-dip turbidites. Whilst soft-sed-

iment deformation features are pervasive in the falling

stage shoreface strata of the Ferron Sandstone,

slumped sandstones that are genetically related to

deposition of (lowstand) turbidites are absent. A sim-

ilar paucity of deformed deposits has been noted from

hyperpycnal turbidites in the Spitsbergen Central

Basin (Plink-Bjorklund and Steel, 2004) and on the

seaward slopes of fjord deltas (Bornhold and Prior,

1990; Prior and Bornhold, 1990) and is argued to

reflect deposition by a hyperpycnal mechanism rather

than slope failure. Delta front facies (e.g., mouth bar

sandstones) are also absent in the lower Ferron stra-

tigraphy suggesting a direct connection between low-

stand rivers and the deeper basin.

3.1.1.3. Sediment bypass in proximal zones. The

presence of surfaces of sediment bypass in proximal

settings with down-dip thickening of turbidite facies is

analogous to the btype 1Q clinoforms described from

the Spitsbergen Central Basin (Plink-Bjorklund and

Steel, 2004). These clinoforms are characterised by

bypass from the shelf edge to the lower slope and

thick accumulations of turbidites from the base of

slope out onto the basin floor (Plink-Bjorklund and

Steel, 2004). Turbidite systems associated with these

clinoforms are associated with river systems that

debouched into channels (chutes) that cut into the

shelf edge and slope (Steel et al., 2000). Sand is

interpreted to have been deposited on the slope by

hyperpycnal flows (Plink-Bjorklund and Steel, 2004).

In the lower Ferron deposits, a similar arrangement of

facies is associated with up-dip erosion and bypass to

down-dip depositional thickening of turbidite beds.

Reconstructing palaeogeography (Fig. 16) illustrates

that the zone of sediment bypass, as represented by

thin gravel deposits of facies association E in areas to

the north and extensive bioturbation into underlying

falling stage deposits, occurs across the Farnham

Dome and Mounds Anticline. Both of these structures

are reported to have been present early in the evolu-

tion of the basin (Molenaar and Cobban, 1991; Wit-

kind, 1988). Ignitive flows are assumed to have

accelerated down such a slope. Whilst elevated depo-

sitional slopes are not a prerequisite for the initiation

of hyperpycnal flows, Plink-Bjorklund and Steel

(2004) argue that sediment transport beyond the

slope is more probable if a direct connection to a

river mouth exits. Furthermore, Prior and Bornhold

(1990) proposed that hyperpycnal flows at delta fronts

transporting sands and gravels are likely where the

bottom surface gradient of nearshore areas is greater

than the slope of the river thalweg. To date, channels

equivalent to those described in the Spitsbergen exam-

ples are not observable. The presence of inferred

topography may account for down-dip differential

thickness variability of the falling stage shoreface

strata (e.g., Section B, Fig. 12) and may have promot-

ed slumping and soft-sediment deformation within

these older units.

3.1.1.4. Change in dominant shoreline depositional

process. Highstand and falling stage shorelines of

the lower Ferron were dominated by storm and wave

processes (e.g., facies associations C and D), resulting

in the transfer of fine-grained sand away from river

mouths by longshore processes (e.g., Bhattacharya

Fig. 16. Palaeogeographic reconstructions for the (A) lower Ferron and (B) Juana Lopez Members. (A) The lowstand systems tract is

represented by turbidite deposition far into the basin, long distances from the underlying falling stage shoreface extent (shown). A submarine

channel is extrapolated northwards towards the Farnham Dome and Mounds area. These structures may have contributed a localised slope down

which flow accelerations took place, causing the turbidity flows to become ignitive. (B) Geostrophic flows transported and deposited Juana

Lopez sands to the east of the San Rafael Anticline from a source area to the north. Their absence to the west may reflect interaction with a

basinfloor structure in the position of the San Rafael Anticline. The inferred flow deflections are recorded by divergent palaeocurrent readings

between sole marks and cross-strata.


and Giosan, 2003). By contrast, lowstand palaeogeo-

graphy was characterised by offshore sediment trans-

port with minimal coast-parallel sand transport.

Further, the transport of coarser material during the

lowstand interval implies a change in the hydrody-

namic regime, perhaps driven by external factors (e.g.,

hinterland tectonics, rate and magnitude of eustatic

sea-level fall, climate) enabling lowstand rivers to

carry greater concentrations of sediment and to cross

the density threshold required for the generation of

riverine underflows. The inception of riverine under-

flows restricted the lateral supply of sediment to

alongshore areas, resulting in a period of sediment

condensation in areas away from the sediment source

area. The interpreted change in depositional process is

consistent with fluid composition histories for

cements in the lower Ferron shoreface sandstones.

Dilution of saline pore fluids during the precipitation

of calcite-cemented concretions in the uppermost part

of the lower Ferron succession (McBride et al., 2003)

therefore probably reflects greater freshwater runoff

into nearshore areas of a relatively constricted low-

stand sea.

3.1.1.5. Flow channelisation. Models of turbidite

channel inception have centred upon progressive ero-

sion by several events (Clark and Pickering, 1996;

Imran et al., 1998) or by the transport and basal scour

by large glide blocks (Elliott, 2000). However, recent

laboratory experiments and numerical models have

shown that the very nature of continuous flows may

result in the self-generation of channels (Imran et al.,

1998; Alexander and Mulder, 2002). These models

illustrate the lateral distribution of net erosion and


deposition as underflows exit from a confined point

source (e.g., a submarine canyon). As the flow exits

from its confined pathway, it expands resulting in the

lateral diminution of erosion away from the centre of

the flow towards the transverse edges. Hence the

inception of these channels could conceivably be a

result of self-channelisation by hyperpycnal flows

exiting a rivermouth point source. Additionally, chan-

nel construction would have produced marked varia-

tions in cross-sectional flow velocities inside and

outside of the channel, producing a variety of over-

bank processes and transporting fine-grained material

over the channel margins.

3.1.1.6. Palaeogeographic considerations. It is well

established that the textural properties of turbidites are

primarily controlled by the nature of the sourcing

fluvial feeder system (Reading, 1991). The unusually

coarse grain-size of facies association E must be a

product of a coarse-grained, possibly braided, river

system transporting sediment to a braid or fan delta.

This supply of sediment cannot have been derived by

local erosion of underlying fine-grained sands of the

falling stage systems tract shoreface deposits.

Schwans (1995) correlated lower Ferron shoreface

strata to braided fluvial deposits in the subsurface of

the Wasatch Plateau, the occurrence of which was

attributed to tectonism and source area uplift along

central Utah salients of the Sevier thrust belt. It is

likely that these coarse-grained fluvial sandstones

belong to the highstand systems tract, deposited dur-

ing accommodation creation in alluvial plain settings.

Subsequent lowering of base-level would have

resulted in steeper fluvial profiles and delivery of

even coarser grained sediment to the shoreline with

a low preservation potential (Posamentier and Morris,

2000). The enhanced carrying capacity of these coars-

er-grained feeder systems would have promoted

hyperpycnal underflows at the shoreline. Submarine

imaging of coarse-grained deltas by sidescan sonar

methods shows them to be characterised by a spec-

trum of delta front chutes and channels, over which a

wide variety of frequently-occurring, long-distance

gravity flows are expected to occur, often initiated

by riverine underflows from the transport of high

sediment loads (Prior and Bornhold, 1989, 1990;

Bornhold and Prior, 1990). Unlike the envisaged

palaeogeography of the lower Ferron, many Late

Cretaceous river systems of other stratigraphic inter-

vals in central Utah were significantly finer-grained

and fed lobate, wave dominated deltas (e.g., Balsley,

1980; Van Wagoner et al., 1990; Kamola and Van

Wagoner, 1995; O’Byrne and Flint, 1995; Van Wag-

oner, 1995). As such they were likely to have been of

significantly lower gradient and sediment yield, and

therefore, less likely to turn hyperpycnal.

3.1.2. Juana Lopez member depositional model

Although the sandstones of the Juana Lopez Mem-

ber are interpreted to have been transported and de-

posited by waning turbulent flows, interpreting the

flow initiation mechanism is somewhat more difficult,

primarily because of the shortage of information on

contemporaneous shoreface environments and archi-

tecture. Secondly, in contrast to the lower Ferron

Sandstone, coarse-grained sandstones such as those

of facies association E have not been observed and

hence there is no evidence for channelisation of the

flows. Thus the sheet-like geometry of these sand-

stones can be feasibly attributed to either turbidity or

geostrophic (wind-driven) currents. Our interpretation

of these deposits is largely dependent on parasequence

stacking patterns, regional stratigraphic considerations

and sea-level context from existing studies.

Turbidity currents resulting from either slope col-

lapse or hyperpycnal flows is considered unlikely for

the Juana Lopez Member because of the relatively

high accommodation setting following transgression

at the end of lower Ferron deposition. The invoked

landward shift in the shoreline would have resulted in

a shallowing of fluvial gradients and sediment being

trapped within up-dip shorefaces and coastal plain

settings. An alternative and favoured hypothesis is

that the Juana Lopez sandstones may be considered

as the deposits of shoreline-oblique geostrophic cur-

rents, generated when winds blow at high angles to

the coast, setting up a pressure gradient between

surface waters moving landward and seaward-return-

ing bottom currents (Duke, 1990). The southerly

palaeoflow directions of Juana Lopez flows is sub-

parallel to oblique to the regional coastline trend in

central Utah (~north–south e.g., Roberts and Kirsch-

baum, 1995). South to southeastward returning geo-

strophic flows are plausible since such processes are

likely to be favoured during times of increased basin

width and wave fetch following transgression and


subsequent highstand. Importantly, a counter-clock-

wise geostrophic circulation pattern is consistent

with computed Turonian oceanographic models for

the Western Interior (Ericksen and Slingerland,

1990; Slingerland et al., 1996). Thus these sandstones

are interpreted as the products of storm-related, shelf

transport processes during sea-level highstand and

coincident with highstand deposition and pronounced

seaward stepping of the upper Ferron deltaic complex

in southern Castle Valley (Gardner, 1995).

3.1.3. Facies distributions

Sandstone facies of the Juana Lopez Member de-

scribed here from outcrops on the eastern side of the

SRA are lacking along its western flank (Molenaar

and Cobban, 1991). The angular variance between

sole marks at the bases of beds and the overlying

current-ripples is conceivably a consequence of inter-

action between incident flows and a gently-dipping

lateral sea-floor structure in the position of the present

day SRA. Subaqueous flow deflections have been

commonly recognised in the ancient record (e.g.,

Fig. 17. Synoptic model based on the lower Ferron Sandstone illustrating th

architecture. Falling stage strata reaching an inclined sea-floor gradien

physiographic continental slope. In such cases, the elevated gradient and

deposition. Hence lowstands in basins lacking a shelf edge may be repres

and coastal plain strata as otherwise predicted by existing sequence stratig

Kelling, 1964; van Andel and Komar, 1969; Ricci

Lucchi and Valmori, 1980; Pickering and Hiscott,

1985) and have been replicated in laboratory flume-

tank experiments (Pantin and Leeder, 1987; Kneller et

al., 1991; Edwards et al., 1994). Distribution of Juana

Lopez sandstones and palaeocurrent variance implies

that the proto-SRA may have exerted a local positive

palaeobathymetric expression that was of sufficient

relief to prevent south-flowing turbidity currents

from reaching the western flank (Fig. 16B).

The proposed physiography exerted by the Farn-

ham Dome northern extension of the SRA during

deposition of the lower Ferron Sandstone, combined

with palaeocurrent variance parallel to the main axis

of the SRA and the partitioning of Juana Lopez

turbidite facies across structure reinforce models that

propose the presence of intra-basinal antecedent phys-

iography inherited prior to foreland basin develop-

ment and/or pre-Tertiary tectonic uplift of the SRA.

In summary, the sandstones belonging to the lower

Ferron Sandstone that are encased by Mancos Shale in

central Utah are considered the products of turbidity

e potential effect of basin floor topography on lowstand depositional

t may operate akin to a shelf edge delta in basins possessing a

raised sediment loads carried by lowstand rivers promotes turbidite

ented by turbidite fans rather than aggradational stacks of shoreface

raphic models (e.g. Plint and Nummedal, 2000).


currents that most likely originated as hyperpycnal

underflows exiting a coarse-grained delta front. The

inception of these turbidites is proposed to be a func-

tion of both high sediment loads reaching a braid delta

(in the subsurface) creating negatively buoyant under-

flows as a result of delta progradation to a structural-

ly-generated, transient shelf edge (Fig. 17). Flow

enhancement through scours or channels enabled the

long-distance transport and eventual deposition of

coarse sands over large distances.

4. Conclusions

The Turonian of central Utah was a site of signif-

icant cross-shelf sediment transport. The mechanisms

by which these transport processes were initiated are

inherently related to their relative sea-level context.

Deposits of the lower Ferron Sandstone Member of

central Utah represent a period of shoreline regression

into the Western Interior Seaway under highstand and

falling sea-level conditions. Lowstand deposits are

represented by turbiditic sandstones that are inter-

preted to be the products of sustained, quasi-steady

flows. The proposed hyperpycnal origin is based upon

the following criteria:

1. Turbidite facies consisting of coarse-grained, cross-

bedded sandstones require sustained flow for their

development.

2. An absence of related slump, debris flow and delta

front facies suggesting a direct connection between

lowstand rivers and the deeper basin.

3. A surface of sediment bypass above falling stage

systems tract strata.

4. A change in dominant depositional shoreline

process from oscillatory during falling sea-level

to offshore sediment transport during sea-level

lowstand.

5. Correlations with coarse-grained fluvial deposits in

the subsurface.

Hyperpycnal flows are considered to have been pro-

moted by enhanced sediment supply during sea-level

lowstand and by structurally-generated sea-floor

topography.

In contrast, sandstones of the overlying Juana

Lopez Member are interpreted to have been deposited

by long-distance, southward-flowing geostrophic cur-

rents following sea-level rise and inundation of the

lower Ferron Sandstone clastic wedge. Progradational

stacking of parasequences of this succession is inter-

preted as deposition under sea-level highstand condi-

tions that is coeval with progradation of a major

deltaic complex in southern Castle Valley. Palaeocur-

rent information and facies mapping indicates the

presence of active sea-floor structures in the area of

the San Rafael Anticline.

The propensity of depositional systems to deliver

sand-grade material over large distances across the

shelf has important ramifications for sequence strati-

graphic models. Given suitable conditions, sea-level

falls in basins lacking a regional shelf edge may

produce lowstand turbidite successions provided

river systems are capable of transporting sufficient

concentrations of sediment to exceed the density

threshold of the ambient (marine) water. Sea-floor

roughness exerted by intra-basinal structures may en-

courage the generation of hyperpycnal flows, contrib-

uting an important modification to generalized models

of forced regression.

Acknowledgements

This paper forms a part of a PhD thesis carried out

by the primary author under the guidance of S. Flint

(University of Liverpool) and J. Howell (University of

Bergen) and funded by a studentship provided jointly

by the University of Liverpool and ENI-LTE. In ad-

dition, the British Sedimentological Research Group

is acknowledged for providing additional financial

support for fieldwork via the Farrell Funding Scheme.

The manuscript benefited from critical reviews pro-

vided by E. Cotter (Bucknell University), R. Fitzsim-

mons (ConocoPhillips, Norway), P. Plink-Bjorklund

(University of Gothenburg) and D. Tabet (Utah Geo-

logical Survey). T. Elliott (University of Liverpool),

R. Gawthorpe (University of Manchester), and W.

Nemec (University of Bergen) are also thanked for

their in-depth and thought-provoking discussions.

References

Alexander, J., Mulder, T., 2002. Experimental quasi-steady density

currents. Marine Geology 186, 195–210.


Balsley, J.K., 1980. Cretaceous Wave-Dominated Delta Systems:

Book Cliffs, East-Central Utah. Amoco Production Company,

Denver, Colorado. 163 pp.

Bates, C.C., 1953. Rational theory of delta formation. American

Association of Petroleum Geologists Bulletin 37, 2119–2162.

Bergman, K.M., 1994. Shannon Sandstone in Hartzog Draw–Heldt

draw fields (Cretaceous, Wyoming, USA) reinterpreted as low-

stand shoreface deposits. Journal of Sedimentary Petrology B64,

184–201.

Bergman, K.M., Walker, R.G., 1987. The importance of sea-level

fluctuations in the formation of linear conglomerate bodies;

Carrot Creek Member of the Cardium Formation, Cretaceous

western interior seaway, Alberta, Canada. Journal of Sedimen-

tary Petrology 57, 651–665.

Bhattacharya, J., Giosan, L., 2003. Wave-influenced deltas: geo-

morphological implications for facies reconstruction. Sedimen-

tology 50, 187–210.

Bornhold, B.D., Prior, D.B., 1990. Morphology and sedimentary

processes on the subaqueous Noeick River delta, British Colum-

bia, Canada. In: Colella, A., Prior, D.B. (Eds.), Coarse-grained

Deltas. International Association of Sedimentologists, vol. 10.

Blackwell Scientific Publications, Oxford, pp. 169–181.

Bottjer, D.J., Arthur, M.A., Dean, W.E., Hattin, D.E., Savrda, C.E.,

1986. Rhythmic bedding produced in Cretaceous pelagic car-

bonate environments: sensitive recorders of climatic cycles.

Paleoceanography 1, 467–481.

Bouma, A.H., 1962. Sedimentology of Some Flysch Deposits: A

Graphic Approach to Facies Interpretation. Elsevier, New York.

168 pp.

Bowen, A.J., Normark, W.R., Piper, D.J.W., 1984. Modelling of

turbidity currents on Navy submarine fan, California borderland.


Brenchley, P.J., 1985. Storm influenced sandstone beds. Modern

Geology 9, 369–396.

Brenchley, P.J., Romano, M., Guiterrez Marco, J.C., 1986. Proximal

and distal hummocky cross-stratified facies on a wide Ordovi-

cian shelf in Iberia. In: Knight, R.J., McLean, J.R. (Eds.), Shelf

Sands and Sandstones. Canadian Society of Petroleum Geolo-

gists, Memoir, vol. 11, pp. 241–256.

Bromley, R.G., 1996. Trace Fossils: Biology, Taphonomy and

Applications. Chapman & Hall, London. 361 pp.

Chan, M.A., Newman, S.L., May, F.E., 1991. Deltaic and shelf

deposits in the Cretaceous Blackhawk Formation and Mancos

Shale, Grand County, Utah. Miscellaneous Publication, 91–6.

Utah Geological Survey. 83 pp.

Chao, S.-Y., 1998. Hyperpycnal and buoyant plumes from a sediment

laden river. Journal of Geophysical Research 103, 3067–3081.

Cheel, R.J., Leckie, D.A., 1993. Hummocky cross stratification. In:

Wright, V.P. (Ed.), Sedimentology Review, vol. 1. Blackwell

Scientific Publications, Oxford, U.K., pp. 103–122.

Chikita, K., 1990. Sedimentation by river-induced turbidity cur-

rents: field measurements and interpretation. Sedimentology

37, 891–905.

Clark, J.D., Pickering, K.T., 1996. Submarine Channels: Processes

and Architecture. Vallis Press, London. 231 pp.

Cole, R.D., Young, R.G., 1991. Facies characterization and archi-

tecture of a muddy shelf-sandstone complex: bMancos BQ inter-

val of Upper Cretaceous Mancos Shale, Northwest Colorado-

Northeast Utah. In: Miall, A.D., Tyler, N. (Eds.), The Three-

Dimensional Facies Architecture of Terrigenous Clastic Sedi-

ments and Its Implications for Hydrocarbon Discovery and

Recovery. Concepts in Sedimentology and Paleontology, vol.

3. SEPM, pp. 277–287.

Cole, R.D., Young, R.G., Willis, G.C., 1997. The Prairie Canyon

Member, a new unit of the Upper Cretaceous Mancos Shale,

west-central Colorado and east-central Utah. Miscellaneous

Publication, 97-4. Utah Geological Survey. 23 pp.

Cotter, E., 1975. Late Cretaceous sedimentation in a low energy

coastal zone: the Ferron Sandstone of Utah. Journal of Sedi-

mentary Petrology 45, 15–41.

Damuth, J.E., Flood, R.D., Kowsman, R.O., Belderson, R.H., Gor-

ini, M.A., 1988. Anatomy and growth pattern of Amazon deep-

sea fan as revealed by long-range side scan sonar (GLORIA)

and high-resolution seismic studies. American Association of

Petroleum Geologists Bulletin 72, 885–911.

DeCelles, P.G., Lawton, T.F., Mitra, G., 1995. Thrust timing, growth

of structural culminations, and synorogenic sedimentation in the

type Sevier Orogenic Belt, western United States. Geology 23

(8), 699–702.

Dott, R.H., Bourgeois, J., 1982. Hummocky stratification: signifi-

cance of its variable bedding sequences. Geological Society of

America Bulletin 93, 663–680.

Duke, W.L., 1990. Geostrophic circulation or shallow marine tur-

bidity currents? The dilemma of paleoflow patterns in storm

influenced prograding shoreline systems. Journal of Sedimenta-

ry Petrology 60, 870–883.

Eaton, J.G., Nations, J.D., 1991. Introduction; tectonic setting along

the margin of the Cretaceous Western Interior Seaway, south-

western Utah and northern Arizona. In: Eaton, J.G., Nations,

J.D. (Eds.), Stratigraphy, Depositional Environments, and Sed-

imentary Tectonics of the Western Margin, Cretaceous Western

Interior Seaway. Special Paper, vol. 260. Geological Society of

America, pp. 1–8.

Eaton, J.G., Kirkland, J.I., Kauffman, E.G., 1990. Evidence and

dating of Mid-Cretaceous tectonic activity in the San Rafael

Swell, Emery County, Utah. The Green Mountain Geologist 27

(2), 39–45.

Edwards, D., Leeder, M.R., Best, J.L., Pantin, H.M., 1994. On

experimental reflected density currents and the interpretation

of certain turbidites. Sedimentology 41, 437–461.

Elliott, T., 2000. Megaflute erosion surfaces and the initiation of

turbidite channels. Geology 28, 119–122.

Ericksen, M.C., Slingerland, R., 1990. Numerical simulations of

tidal and wind-driven circulation in the Cretaceous Interior

Seaway of North America. Geological Society of America

Bulletin 102, 1499–1516.

Franczyk, K.J., Pitman, J.K., 1991. Latest Cretaceous non-marine

depositional systems in the Wasatch Plateau area: reflections of

foreland to intermontane basin transition. In: Chidsey, T.C.

(Ed.), Geology of East-Central Utah. Utah Geological Associa-

tion Publication, vol. 19, pp. 77–94.

Gardner, M.H., 1995. The stratigraphic hierarchy and tectonic

history of the mid-Cretaceous foreland basin of Central

Utah. In: Dorobek, S.L., Ross, G.M. (Eds.), Stratigraphic Evo-


lution of Foreland Basins, Special Publication, vol. 52. SEPM,

pp. 283–303.

Hale, L.A., 1972. Depositional history of the Ferron Formation,

central Utah, in Plateau Basin and Range Transition Zone. Utah

Geological Association, Salt Lake City, pp. 115–138.

Hale, L.A., Van de Graaff, F.R., 1964. Cretaceous stratigraphy and

facies patterns — northeastern Utah and adjacent areas, geology

and mineral resources of the Uinta Basin. Intermountain Asso-

ciation of Petroleum Geologists 13th Annual Field Guidebook,

Salt Lake City, pp. 115–138.

Hamblin, A.P., Walker, R.G., 1979. Storm-dominated shallow ma-

rine deposits: the Fernie–Kootenay (Jurassic) transition, south-

ern Rocky Mountains. Canadian Journal of Earth Sciences 16,

1673–1690.

Hampson, G.J., Howell, J.A., Flint, S.S., 1999. A sedimentological

and sequence stratigraphic re-interpretation of the Upper Creta-

ceous Prairie Canyon Member (bMancos BQ) and associated

strata, Book Cliffs area, Utah, USA. Journal of Sedimentary

Research 69 (2), 414–433.

Haq, B.U., Hardenbol, J., Vail, P.R., 1988. Mesozoic and Cenozoic

chronostratigraphy and cycles of sea-level change. In: Wilgus,

C.K., et al., (Eds.), Sea-Level Changes: An Integrated Ap-

proach. Special Publication, vol. 42. SEPM, pp. 71–108.

Harms, J.C., Southard, J.B., Walker, R.G., 1982. Structure

and sequence in clastic rocks. SEPM Short Course, vol. 9,

pp. 133–161.

Hay, A.E., 1987. Turbidity currents and submarine channel forma-

tion in Rupert Inlet, British Columbia: 2. The roles of continu-

ous and surge-type flows. Journal of Geophysical Research 92,

2883–2900.

Heller, P.L., Paola, C., 1989. The paradox of lower cretaceous

gravels and the initiation of thrusting in the Sevier Orogenic

Belt, United States Western Interior. Geological Society of

America Bulletin 101, 864–875.

Heller, P.L., Beekman, F., Angevine, C.L., Cloetingh, S.A.P.L.,

1993. Cause of tectonic reactivation and subtle uplifts in the

Rocky Mountain region and its effect on the stratigraphic re-

cord. Geology 21, 1003–1006.

Hettinger, R.D., McCabe, P.J., Shanley, K.W., 1994. Detailed

facies anatomy of transgressive and highstand systems

tracts from the Upper Cretaceous of southern Utah,

USA. In: Weimer, P., Posamentier, H.W. (Eds.), Siliciclastic

Sequence Stratigraphy: Recent Developments and Applications,

American Association of Petroleum Geologists Memoir, vol. 58,

pp. 235–257.

Hicks, D.M., Gomez, B., Trustrum, N.A., 2004. Event suspended

sediment characteristics and the generation of hyperpycnal

plumes at river mouths: East Coast Continental Margin, North

Island, New Zealand. Journal of Geology 112, 471–485.

Hiscott, R.N., Pickering, K.T., Bouma, A.H., Hand, B.M., Kneller,

B.C., Postma, G., Soh, W., 1997. Basin-floor fans in the North

Sea: sequence stratigraphic models vs. sedimentary facies: dis-

cussion. American Association of Petroleum Geologists Bulletin

81, 662–665.

Howell, J.A., Flint, S., Hunt, C.B., 1996. Sedimentological aspects

of the Humber Group (Upper Jurassic) of the South Central

Graben, UK North Sea. Sedimentology 43 (1), 89–114.

Hughes-Clarke, J.E., Shor, A.N., Piper, D.J.W., Mayer, L.A., 1990.

Large-scale, current-induced erosion and deposition in the path

of the 1929 Grand Banks turbidity current. Sedimentology 37,

613–629.

Hunt, D., Tucker, M.E., 1992. Stranded parasequences and the

forced regressive wedge systems tract: deposition during base-

level fall. Sedimentary Geology 81, 1–9.

Imran, J., Parker, G., Katopodes, N., 1998. A numerical model of

channel inception on submarine fans. Journal of Geophysical

Research 103, 1219–1238.

Imran, J., Parker, G., Harff, P., 2002. Experiments on incipient

channelization of submarine fans. Journal of Hydraulic Re-

search 40, 21–32.

Jennette, D.C., Jones, C.R., 1995. Sequence tratigraphy of the

Upper Cretaceous Tocito Sandstone: a model for tidally-influ-

enced incised valleys, San Juan Basin, New Mexico. In: Van

Wagoner, J.C., Bertram, G.T. (Eds.), Sequence Stratigraphy of

Foreland Basin Deposits: Outcrop and Subsurface Examples

from the Cretaceous of North America, American Association

of Petroleum Geologists Memoir, vol. 64. AAPG (American

Association of Petroleum Geologists), Tulsa, pp. 311–347.

Johnson, K.S., Paull, C.K., Barry, J.P., Chavez, F.P., 2001. A decal

record of underflows from a coastal river into the deep sea.

Geology 29, 1019–1022.

Jordan, T.E., 1981. Thrust loads and foreland basin evolution,

Cretaceous, western United States. Geology 22, 1139–1143.

Kamola, D.L., Van Wagoner, J.C., 1995. Stratigraphy and facies

architecture of parasequences with examples from the Spring

Canyon Member, Blackhawk Formation, Utah. In: Van Wagon-

er, J.C., Bertram, G.T. (Eds.), Sequence Stratigraphy of Foreland

Basin Deposits: Outcrop and Subsurface Examples from the

Cretaceous of North America. American Association of Petro-

leum Geologists Memoir, vol. 64, pp. 27–54.

Kassem, A., Imran, J., 2001. Simulation of turbid underflows

generated by the plunging of a river. Geology 29, 655–658.

Kauffman, E.G., 1977. Geological and biological overview: West-

ern Interior Cretaceous Basin. The Green Mountain Geologist

14 (3–4), 75–99.

Kauffman, E.G., 1984. Paleobiogeography and evolutionary re-

sponse dynamic in the Cretaceous Western Interior Seaway of

North America. In: Westermann, G.E.G. (Ed.), Mesozoic Bio-

geography of North America. Geological Association of Canada

Special Paper, vol. 7, pp. 273–306.

Kauffman, E.G., Sageman, B.B., Kirkland, J.I., Elder, W.P., Harries,

P.J., Villamil, T., 1993. Molluscan biostratigraphy of the Creta-

ceous Western Interior Basin, North America. In: Caldwell,

W.G.E., Kauffman, E.G. (Eds.), Evolution of the Western Inte-

rior Basin. St John’s, Geological Association of Canada Special

Paper, vol. 39, pp. 397–434.

Kelling, G., 1964. The turbidite concept in Britain. In: Bouma,

A.H., Brouwer, A. (Eds.), Turbidites, Developments in Sedi-

mentology, vol. 3. Elsevier, Amsterdam, pp. 75–92.

Kineke, G.C., Woolfe, K.J., Kuehl, S.A., Milliman, J.D.,

Dellapenna, T.M., Purdon, R.G., 2000. Sediment export

from the Sepik River, Papua New Guinea: evidence for

a divergent sediment plume. Continental Shelf Research 20,

2239–2266.


Kneller, B.C., 1995. Beyond the turbidite paradigm: physical mod-

els for deposition of turbidites and their implications for reser-

voir prediction. In: Hartley, A.J., Prosser, D.J. (Eds.),

Characterisation of deep marine clastic systems. Geological

Society of London, pp. 31–49. Special Publication.

Kneller, B.C., Branney, M.J., 1995. Sustained, high-density turbid-

ity currents and the deposition of thick, massive beds. Sedimen-

tology 42, 607–616.

Kneller, B.C., Buckee, C., 2000. The structure and fluid mechan-

ics of turbidity currents: a review of some recent studies and

their geological implications. Sedimentology 47 (Suppl. 1),

62–94.

Kneller, B., Edwards, D., McCaffrey, W., Moore, R., 1991. Oblique

reflection of turbidity currents. Geology 14, 250–252.

Kolla, V., Perlmutter, M.A., 1993. Timing of turbidite sedimentation

on the Mississippi Fan. American Association of Petroleum

Geologists Bulletin 77, 1129–1141.

Lawton, T.F., 1983. Late Cretaceous fluvial systems and the age of

foreland uplifts in central Utah. In: Lowell, J.D., Gries, R.

(Eds.), Rocky Mountain Foreland Basins and Uplifts. Rocky

Mountain Association of Geologists, pp. 181–200.

Lawton, T.F., 1986. Fluvial systems of the Upper Cretaceous Mesa-

verde Group and Paleocene North Horn Formation, central

Utah: a record of transition from thin-skinned to thick-skinned

deformation in the foreland region. In: Peterson, J.A. (Ed.),

Paleotectonics and Sedimentation, American Association of

Petroleum Geologists Memoir, vol. 41. AAPG, Tulsa, Okla-

homa, pp. 423–442.

Leckie, D.A., Krystinick, L.F., 1989. Is there evidence for geo-

strophic currents preserved in the geological record of inner to

middle shelf deposits? Journal of Sedimentary Petrology 59,

862–870.

Leckie, D.A., Walker, R.G., 1982. Storm and tide-dominated

shorelines in Cretaceous Moosebar–Lower Gates interval—

outcrop equivalents of deep basin gas trap in Western Canada.


138–157.

Lowe, D., 1982. Sediment gravity flows: II. Depositional models

with special reference to the deposits of high-density turbidity


Lupton, C.T., 1916. Geology and coal resources of Castle Valley in

Carbon, Emery and Sevier Counties, Utah. United States Geo-

logical Survey Bulletin 628 (84 pp.).

Major, J.J., 1997. Depositional processes in large-scale debris-flow

experiments. Journal of Geology 105, 345–366.

Martinson, V.S., Heller, P.L., Frerichs, W.E., 1998. Distinguish-

ing middle Late Cretaceous tectonic events from regional

sea-level change using foraminiferal data from the US West-

ern Interior. Geological Society of America Bulletin 110 (2),

259–268.

McBride, E.F., Picard, M.D., Milliken, K.L., 2003. Calcite cemen-

ted concretions in Cretaceous sandstone, Wyoming and Utah,

USA. Journal of Sedimentary Research 73, 462–483.

McGookey, D.P., Haun, J.D., Hale, L.A., McCubbin, D.G., Weimer,

R.J., Wulf, G.R., 1972. Cretaceous system. In: Mallory, W.W.

(Ed.), Geologic Atlas of the Rocky Mountain Region. Rocky

Mountain Association of Geologists, Denver, pp. 190–228.

McLeod, P., Carey, S., Sparks, R.S.J., 1999. Behaviour of particle-

laden flows in the ocean: experimental simulation and geolog-

ical implications. Sedimentology 46, 523–537.

Mellere, D., Plink-Bjorklund, P., Steel, R., 2002. Anatomy of shelf

deltas at the edge of a prograding Eocene shelf margin, Spits-

bergen. Sedimentology 49, 1181–1206.

Merewether, E.A., Cobban, W.A., 1986. Biostratigraphic units and

tectonism in the mid-Cretaceous foreland of Wyoming, Color-

ado and adjoining areas. In: Peterson, J.A. (Ed.), Paleotectonics

and Sedimentation. American Association of Petroleum Geolo-

gists Memoir, vol. 41, pp. 443–468.

Molenaar, C.M., Cobban, W.A., 1991. Middle Cretaceous stratig-

raphy on the south and east sides of the Uinta Basin, northeast-

ern Utah and northwestern Colorado. US Geological Survey

Bulletin 1787-P (34 pp.).

Molenaar, C.M., Wilson, B.W., 1990. The Frontier Formation and

associated rocks of northeastern Utah and northwestern Color-

ado. US Geological Survey Bulletin 1787-M (21 pp.).

Montgomery, S.L., Tabet, D.E., Barker, C.E., 2001. Upper Creta-

ceous Ferron Sandstone: major coalbed methane play in central

Utah. American Association of Petroleum Geologists Bulletin

85, 199–219.

Mulder, T., Alexander, J., 2001. The physical character of subaque-

ous sedimentary density flows and their deposits. Sedimentolo-

gy 48, 269–299.

Mulder, T., Syvitski, J.P.M., 1995. Turbidity currents generated at

river mouths during exceptional discharges to the worlds

oceans. Journal of Geology 103, 285–299.

Mulder, T., Syvitski, J.P.M., Skene, K.I., 1998. Modelling of ero-

sion and deposition by turbidity currents generated at river

mouths. Journal of Sedimentary Research 68, 124–137.

Mulder, T., Migeon, S., Savoye, B., Faugeres, J.-C., 2001. Inversely

graded turbidite sequences in the deep Mediterranean: a record

of deposits from flood-generated turbidity currents? Geo-marine

Letters 21, 86–93.

Mulder, T., Syvitski, J.P.M., Migeon, S., Faugeres, J.-C., Savoye,

B., 2003. Marine hyperpycnal flows: initiation, behavior and

related deposits. A review. Marine and Petroleum Geology 20,

861–882.

Mutti, E., 1985. Turbidite systems and their relations to depositional

sequences. In: Zuffa, G.G. (Ed.), Provenance of Arenites. Rei-

del, Dordrecht, pp. 65–93.

Mutti, E., Normark, W.R., 1991. An integrated approach to the study

of turbidite systems. In: Weimer, P., Link, M.H. (Eds.), Seismic

Facies and Sedimentary Processes of Submarine Fans and Tur-

bidite Systems. Springer-Verlag, New York, pp. 75–106.

Mutti, E., Tinterri, R., Benevelli, G., di Biase, D., Cavanna,

G., 2003. Deltaic, mixed and turbidite sedimentation of

ancient foreland basins. Marine and Petroleum Geology 20,

733–755.

Myrow, P.M., Southard, J.B., 1996. Tempestite deposition. Journal

of Sedimentary Research 66, 875–887.

Nemec, W., 1995. The dynamics of deltaic suspension plumes. In:

Oti, M.N., Postma, G. (Eds.), Geology of Deltas. A.A. Balkema

Publishers, Rotterdam, pp. 31–93.

Normark, W.R., Piper, W.R., 1991. Initiation processes and flow

evolution of turbidity currents: implications for the depositional


record. In: Osborne, R.H. (Ed.), From Shoreline to Abyss:

Contribution in Marine Geology in Honor of Francis Parker

Shepard. Special Publication, vol. 46. SEPM, pp. 207–230.

Normark, W.R., Posamentier, H.W., Mutti, E., 1993. Turbidite

systems: state of the art and future directions. Reviews of

Geophysics B31, 91–116.

O’Byrne, C.J., Flint, S., 1995. Sequence, parasequence, and intra-

parasequence architecture of the Grassy Member, Blackhawk

Formation, Book Cliffs, Utah, U.S.A. In: Van Wagoner, J.C.,

Bertram, G.T. (Eds.), Sequence Stratigraphy of Foreland Basin

Deposits: Outcrop and Subsurface Examples from the Creta-

ceous of North America. American Association of Petroleum

Geologists Memoir, vol. 64, pp. 225–256.

Pantin, H.M.P., Leeder, M.R., 1987. Reverse flow in turbidity

currents: the role of internal solutions. Sedimentology 34,

1143–1155.

Parker, G., 1982. Conditions for the ignition of catastrophically

erosive turbidity currents. Marine Geology 46, 302–327.

Parsons, J.D., Bush, J.W.M., Syvitski, J.P.M., 2001. Hyperpycnal

plume formation from riverine outflows with small sediment

concentrations. Sedimentology 48, 465–478.

Patterson, S.A.J., 2005. Storm-influenced prodelta turbidite com-

plex in the Lower Kenilworth Member at Hatch Mesa, Book

Cliffs, Utah, USA: Implications for shallow marine facies mod-

els. Journal of Sedimentary Research 75, 420–439.

Pemberton, S.G., Spila, M., Pulham, A.J., Saunders, T., MacEa-

chern, J.A., Robbins, R., Sinclair, I.K., 2001. Ichnology and

sedimentology of shallow to marginal marine systems: Ben

Nevis and Avalon Reservoirs, Jeanne d’Arc Basin. Short

Course Notes - Geological Association of Canada, vol. 15.

343 pp.

Peterson, F., 1986. Jurassic paleotectonics in the west-central part of

the Colorado Plateau, Utah and Arizona. In: Peterson, J.A.

(Ed.), Paleotectonics and Sedimentation, American Association

of Petroleum Geologists Memoir, vol. 41, pp. 563–595.

Peterson, F., Ryder, R.T., Law, B.E., 1980. Stratigraphy, sedimen-

tology, and regional relationships of the Cretaceous system in

the Henry Mountains region, Utah. In: Picard, M.D. (Ed.),

Henry Mountains Symposium. Utah Geological Association

Publication, vol. 8, pp. 151–170.

Pickering, K.T., Hiscott, R.N., 1985. Contained (reflected) turbidity

currents from the Middle Ordovician Cloridorme Formation,

Quebec, Canada: an alternative to the antidune hypothesis.


Pickering, K.T., Stow, D.A.V., Watson, M.P., Hiscott, R.N., 1986.

Deep-water facies, processes and models: a review and classi-

fication scheme for modern and ancient sediments. Earth-Sci-

ence Reviews 23, 75–174.

Piper, D.J.W., Normark, W.R., 1983. Turbidite depositional patterns

and flow characteristics, Navy submarine fan, California bor-

derland. Sedimentology 30, 681–694.

Piper, D.J., Savoye, B., 1993. Processes of late Quaternary turbidity

current flow and deposition on the Var deep-sea fan, north-west

Mediterranean Sea. Sedimentology 40, 557–582.

Piper, D.J.W., Shor, A.N., Hughes-Clarke, J.E., 1988. The 1929

Grand Banks earthquake, slump and turbidity current. Special

paper - Geological Society of America 229, 77–92.

Piper, D.J., Hiscott, R.N., Normark, W.R., 1999. Outcrop-scale

acoustic facies analysis and latest Quaternary development of

the Hueneme and Dume submarine fans, offshore California.


Plink-Bjorklund, P., Steel, R.J., 2004. Initiation of turbidity cur-

rents: outcrop evidence for Eocene hyperpycnal flow turbidites.

Sedimentary Geology 165, 29–52.

Plint, A.G., 1988. Sharp-based shoreface sequences and ’offshore

bars’ in the Cardium Formation of Alberta: their relationship to

changes in relative sea-level. In: Wilgus, C.K., et al., (Eds.),

Sea-Level Changes: an Integrated Approach, Special Publica-

tion, vol. 42. SEPM, pp. 357–370.

Plint, A.G., Nummedal, D., 2000. The falling stage systems tract:

recognition and importance in stratigraphic analysis. In: Hunt,

D., Gawthorpe, R.L. (Eds.), Sedimentary Responses to Forced

Regressions. Special Publication - Geological Society of Lon-

don, vol. 172, pp. 1–17.

Posamentier, H.W., Morris, W.R., 2000. Aspects of the stratal

architecture of forced regressive deposits. In: Hunt, D.,

Gawthorpe, R.L. (Eds.), Sedimentary Responses to Forced

Regressions. Special Publication - Geological Society of Lon-

don, vol. 172, pp. 19–46.

Posamentier, H.W., Erskine, R.D., Mitchum, R.M., 1991. Models

for submarine fan deposition within a sequence stratigraphic

framework. In: Weimer, P., Link, M.H. (Eds.), Seismic Facies

and Sedimentary Processes of Submarine Fans and Turbidite

Systems. Springer-Verlag, New York, pp. 127–136.

Posamentier, H.W., Allen, G.P., James, D.P., Tesson, M.,

1992. Forced regressions in a sequence stratigraphic frame-

work: concepts, examples and exploration significance.


1687–1709.

Postma, G., Nemec, W., Kleinspehn, K.L., 1988. Large floating

clasts in turbidites: a mechanism for their emplacement. Sedi-

mentary Geology 58, 47–61.

Prior, D.B., Bornhold, B.D., 1989. Submarine sedimentation

on a developing Holocene fan delta. Sedimentology 36,

1053–1076.

Prior, D.B., Bornhold, B.D., 1990. The underwater development of

Holocene fan deltas. In: Colella, A., Prior, D.B. (Eds.), Coarse-

grained deltas, Special publication of the International Associ-

ation of Sedimentologists, vol. 10. Blackwell Scientific Publica-

tions, Oxford, pp. 75–90.

Prior, D.B., Bornhold, B.D., Johns, M.W., 1986. Active sand trans-

port along a fjord-bottom channel, Bute Inlet, British Columbia.

Geology 14, 581–584.

Reading, H.G., 1991. The classification of deep-sea depositional

systems by sediment calibre and feeder systems. Journal of the

Geological Society (London) 148, 427–430.

Ricci Lucchi, F., Valmori, E., 1980. Basinwide turbidites in a

Miocene oversupplied deep sea plain: a geometric analysis.


Riemersma, P., Chan, M.A., 1991. Facies of the lower Ferron

Sandstone and Blue Gate Shale Members of the Mancos

Shale: lowstand and early transgressive facies architecture. In:

Swift, D.J.P., Oertel, G.F., Tillman, R.W., Thorne, J.A. (Eds.),

Shelf Sands and Sandstone Bodies, Geometry, Facies, and


Sequence Stratigraphy, Special publication of the International

Association of Sedimentologists, vol. 14. International Associ-

ation of Sedimentologists, Oxford, pp. 489–510.

Rimoldi, B., Alexander, J., Morris, S., 1996. Experimental turbidity

currents entering density-stratified water: analogues for turbi-

dites in Mediterranean hypersaline basins. Sedimentology 43,

527–540.

Roberts, L.N.R., Kirschbaum, M.A., 1995. Paleogeography of the

late cretaceous of the Western interior of Middle North Amer-

ica—coal distribution and sediment accumulation. U.S. Geolog-

ical Survey Professional Paper 1561 (115 pp.).

Ryer, T.A., 1977. Patterns of Cretaceous shallow-marine sedimen-

tation, Coalville and Rockport areas, Utah. Geological Society

of America Bulletin 88, 177–188.

Ryer, T.A., 1981. Deltaic coals of Ferron Sandstone Member of the

Mancos Shale: predictive model for Cretaceous coal-bearing

strata of Western Interior. American Association of Petroleum

Geologists Bulletin 65, 2323–2340.

Ryer, T.A., 1991. Stratigraphy, facies and depositional history

of the Ferron Sandstone in the canyon of Muddy Creek,

east-central Utah. In: Chidsey, T.C. (Ed.), Geology of East-

Central Utah, Utah Geological Association Publication, vol.

19, pp. 45–54.

Ryer, T.A., Lovekin, J.R., 1986. The Upper Cretaceous Vernal Delta

of Utah— depositional or paleotectonic feature. In: Peterson,

J.A. (Ed.), Paleotectonics and Sedimentation in the rocky moun-

tain region, United States; Part III. American Association of

Petroleum Geologists Memoir, vol. 41, pp. 497–509.

Ryer, T.A., McPhillips, M., 1983. Early Late Cretaceous paleoge-

ography of east-central Utah. In: Reynolds, M.W., Dolly, E.D.

(Eds.), Mesozoic paleogeography of West-Central United States,

Rocky Mountain Paleogeography Symposium, vol. 2. SEPM,

Rocky Mountain Section, pp. 253–271.

Savrda, C.E., Bottjer, D.J., Seilacher, A., 1991. Redox-related ben-

thic events. In: Einsele, G., Ricken, W., Seilacher, A. (Eds.),

Cycles and Events in Stratigraphy. Springer-Verlag, Berlin,

pp. 524–541.

Schwans, P., 1995. Controls on sequence stacking and fluvial to

shallow marine architecture in a foreland basin. In: Van Wag-

oner, J.C., Bertram, G.T. (Eds.), Sequence Stratigraphy of Fore-

land Basin Deposits: Outcrop and Subsurface Examples from

the Cretaceous of North America. American Association of

Petroleum Geologists Memoir, vol. 64, pp. 55–102.

Schwartz, R.K., DeCelles, P.G., 1988. Cordilleran foreland basin

evolution in response to interactive Cretaceous thrusting and

foreland partitioning, southwestern Montana. In: Schmidt, C.J.,

Perry, W.J. (Eds.), Interaction of the Rocky Mountain Foreland

and the Cordilleran Thrust Belt. Geological Society of America,

Memoir, vol. 171, pp. 489–513.

Shanmugam, G., 1996. High-density turbidity currents: are they

sandy debris-flows? Journal of Sedimentary Research A66,

2–10.

Shanmugam, G., Bloch, R.B., Mitchell, S.M., Beamish, B.W.J.,

Hodgkinson, R.J., Damuth, J.E., Straume, T., Syvertsen, S.E.,

Shields, K.E., 1995. Basin floor fans in the North Sea: sequence

stratigraphic vs. sedimentary facies. American Association of

Petroleum Geologists Bulletin 79, 477–512.

Slingerland, R.L., Kump, L.R., Arthur, M.A., Fawcett, P.J., Sage-

man, B.B., Barron, E.J., 1996. Estuarine circulation in the

Turonian Western Interior seaway of North America. Geological

Society of America Bulletin 108 (7), 941–952.

Southard, J.B., Lambie, J.M., Federico, D.C., Pile, H.T., Weidman,

C.R., 1990. Experiments on bed configurations in fine sand

under bidirectional purely oscillatory flow, and the origin of

hummocky cross stratification. Journal of Sedimentary Petrolo-

gy 60, 1–17.

Sparks, R.S.J., Bonnecaze, R.T., Huppert, H.E., Lister, J.R., Hall-

worth, M.A., Mader, H., Phillips, J., 1993. Sediment-laden

gravity currents with reversing buoyancy. Earth and Planetary

Science Letters 114, 243–257.

Spieker, E.M., Reeside, J.B., 1925. Cretaceous and Tertiary forma-

tions of the Wasatch Plateau, Utah. Geological Society of

America Bulletin vol. 36, 435–454.

Steel, R.J., Crabaugh, J., Schellpeper, M., Mellere, D., Plink-

Bjorklund, P., Deibert, J., Loeseth, T., 2000. Deltas vs.

rivers on the shelf edge: their relative contributions to the

growth of shelf-margins and basin-floor fans (Barremian

and Eocene, Spitsbergen). GCSSEPM Foundation 20th Annu-

al Research Conference: Deep-Water Reservoirs of the World,

pp. 981–1009.

Stephenson, R.A., Cloetingh, S.A.P.L., 1991. Some examples and

mechanical aspects of continental lithospheric folding. Tectono-

physics 188, 27–37.

Stevens, K.M., Chaiwongsaen, N., 2003. Evaluating the origin of

isolated sandstones encased in marine mudstone: the Cretaceous

Prairie Canyon Member of the Mancos Shale, Hatch Mesa,

East-Central Utah. American Association of Petroleum Geolo-

gists Annual Meeting Conference Program and Abstracts, Salt

Lake City, vol. A163.

Swift, D.J.P., Rice, D.D., 1984. Sand bodies on muddy shelves:

a model for sedimentation in the Western Interior Creta-

ceous Seaway. In: Tillman, R.W., Siemers, C.T. (Eds.), Silici-

clastic Shelf Sediments. Special Publication, vol. 34. SEPM,

Tulsa, pp. 25–37.

Swift, D.J.P., Hudelson, P.M., Brenner, R.L., Thompson, P., 1987.

Shelf construction in a foreland basin: storm beds, shelf sand-

bodies and shelf-slope depositional sequences in the Upper

Cretaceous Mesaverde Group, Book Cliffs, Utah. Sedimentolo-

gy 34, 423–457.

Taylor, A.M., Goldring, R., 1993. Description and analysis of

bioturbation and ichnofabric. Journal of Geological Society of

London 150, 141–148.

Taylor, D.R., Lovell, R.W.W., 1995. High frequency sequence

stratigraphy and paleogeography of the Kenilworth Member,

Blackhawk Formation, Book Cliffs, Utah, USA. In: Van Wag-

oner, J.C., Bertram, G.T. (Eds.), Sequence Strtigraphy of Fore-

land Basin Deposits: Outcrop and Subsurface examples from the

Cretaceous of North America. American Association of Petro-

leum Geologists Memoir, vol. 64, pp. 257–276.

Tillman, R.W., Martinsen, R.S., 1987. Sedimentologic model and

production characteristics of Hartzog Draw field, Wyoming, a

Shannon shelf ridge sandstone. In: Tillman, R.W., Weber, K.J.

(Eds.), Reservoir Sedimentology. Special Publication, vol. 40.

SEPM, Tulsa, pp. 15–112.


van Andel, T.H., Komar, P.D., 1969. Ponded sediments of the Mid-

Atlantic Ridge between 228 and 238 north latitude. Geological

Society of America Bulletin 80, 1163–1190.

Van Wagoner, J.C., 1995. Sequence stratigraphy and marine to non-

marine facies architecture of foreland basin strata, Book Cliffs,

Utah, USA. In: Van Wagoner, J.C., Bertram, G.T. (Eds.), Se-

quence Stratigraphy of Foreland Basin Deposits: Outcrop and

Subsurface Examples from the Cretaceous of North America.

American Association of Petroleum Geologists Memoir, vol. 64,

pp. 137–224.

Van Wagoner, J.C., Posamentier, H.W., Mitchum, R.M., Vail,

P., Sarg, J.F., Loutit, T.S., Hardenbol, J., 1988. An

overview of the fundamentals of sequence stratigraphy

and key definitions. In: Wilgus, C.K., et al., (Eds.), Sea-

Level Changes: An Integrated Approach. Special Publication,

vol. 42. SEPM, pp. 39–45.

Van Wagoner, J.C., Mitchum, R.M., Campion, K.M., Rahmanian,

V.D., 1990. Siliciclastic sequence stratigraphy in well logs,

cores, and outcrops: concepts for high-resolution correlation of

time and facies. Methods in Exploration Series, vol. 7. Ameri-

can Association of Petroleum Geologists. 55 pp.

Walker, R.G., 1984. Shelf and shallow marine sands. In: Walker,

R.G. (Ed.), Facies Models. Geoscience Canada Reprint Series,

vol. 1, pp. 141–170. St John’s.

Walker, R.G., Bergman, K.M., 1993. Shannon Sandstone in Wyom-

ing: a shelf ridge complex reinterpreted as lowstand shoreface

deposits. Journal of Sedimentary Petrology 63, 839–851.

Walker, R.G., Plint, A.G., 1992. Wave-and storm-dominated shal-

low marine systems. In: Walker, R.G., James, N.P. (Eds.), Facies

Models: Response to Sea Level Change. Geological Association

of Canada, Waterloo, Ontario, pp. 219–238.

Weimer, R.J., 1962. Late Jurassic and Early Cretaceous correla-

tions, south-central Wyoming and northwestern Colorado.

Wyoming Geological Association 17th Annual Field Confer-

ence, pp. 124–130.

Weimer, P., 1991. Seismic facies, characteristics and variations in

channel evolution, Mississippi Fan (Plio–Pleistocene), Gulf of

Mexico. In: Weimer, P., Link, M.H. (Eds.), Seismic Facies and

Sedimentary Processes of Submarine Fans and Turbidite Sys-

tems. Springer-Verlag, New York, pp. 323–347.

Weirich, F., 1986. The record of density-induced under-flows in a

glacial lake. Sedimentology 33, 261–277.

Williams, G.D., Stelck, C.R., 1975. Speculations on the Creta-

ceous paleogeography of North America. In: Caldwell,

W.G.E. (Ed.), The Cretaceous System in the Western Interior

of North America, Special Paper - Geological Association of

Canada, vol. 13. Geological Association of Canada, Calgary,

pp. 1–20.

Wiltschko, D.B., Dorr, J.A., 1983. Timing of deformation in

overthrust belt and foreland of Idaho, Wyoming and Utah.


1304–1332.

Witkind, I.J., 1988. Geologic map of the Huntington 30�60 quad-

rangle, Carbon, Emery, Grand and Uintah counties, Utah. USGS

Miscellaneous Investigations Series I-1764.

Wright, J.D., Wiseman, W.J., Bornhold, B.D., Prior, D.B., Suhayda,

J.N., Keller, G.H., Yang, Y.-S., Fan, Y.B., 1988. Marine dis-

persal and deposition of Yellow River silts by gravity-driven

underflows. Nature 332, 629–632.

Zeng, J., Lowe, D.R., Prior, D.B., Wiseman, W.J., Bornhold, B.D.,

1991. Flow properties of turbidity currents in Bute Inlet, British

Columbia. Sedimentology 38, 975–996.



Imaging bed geometry and architecture of massive sandstones

in the Fontanelice Channels, Italian Apennines,

using new digiscoping techniques

R.B. Wynn a,*, P.J. Talling b, L. Amy b

aNational Oceanography Centre, European Way, Southampton, SO14 3ZH, UKbDepartment of Earth Sciences, University of Bristol, Queens Road, Bristol, BS8 1RJ, UK

Abstract

In this study we present digital images and sedimentological data from a channel fill succession in the Italian Apennines that

is dominated by massive sandstones. Although the studied outcrop is largely inaccessible, valuable data have now been

obtained using the new technique of ddigiscopingT, which allows features of b10 cm to be resolved from a distance of several

hundred metres.

About 75–80% of the channel fill is composed of massive sandstone beds N1 m thick, with overall sandstone : shale ratios of

~9 :1. Massive sandstones are poorly sorted and overall show little or no normal grading. They are commonly amalgamated and

always have sharp bed tops. Massive sandstone beds show abrupt pinch-outs at the channel margin, whereas overlying thin-

bedded siltstone/mudstone layers taper gradually and drape up the margin more extensively. This suggests that the depositing

flows were stratified into a lower, thin, (hyper)concentrated density flow and an upper, more dilute, turbidity current. In

summary, the digiscoping technique is shown to be a cheap and efficient method for imaging distant and/or inaccessible

outcrops and providing information on bed geometry and architecture.


Keywords: Channel; Marnoso Arenacea; Massive sandstone; Digiscoping

1. Introduction

Deep-water gravity flow processes and deposits

have been the focus of intensive research over the


doi:10.1016/j.sedgeo.2005.04.012

* Corresponding author. Tel.: +44 2380 596553; fax: +44 2380

596554.

E-mail address: [email protected] (R.B. Wynn).

last three or four decades (see recent overviews by

Bouma and Stone, 2000; Shanmugam, 2000; Knel-

ler and Buckee, 2000; Mulder and Alexander,

2001). However, our inability to directly observe

and instrument such flows means that knowledge

of flow processes, and their impact on deposit

character, is restricted to interpretation of preserved

deposits and experimental/numerical modelling.

9 (2005) 153–162

R.B. Wynn et al. / Sedimentary Geology 179 (2005) 153–162154

Such interpretations are inevitably ambiguous and

frequently controversial; this is exemplified by the

continuing debate over the processes responsible for

depositing thick massive sandstones in deep-water

environments (e.g. Lowe, 1982; Kneller and Bran-

ney, 1995; Shanmugam et al., 1995, 1997; Shanmu-

gam, 1996, 2000, 2002; Hiscott et al., 1997; Stow

and Johansson, 2000; Marr et al., 2001). As a

consequence, there is a growing requirement for

new case studies of massive sandstones in both

modern and ancient environments.

In this study we use a new technique called

digiscoping to image the geometry and architecture

of thick massive sandstone beds in a channel fill

succession in the Italian Apennines. Digital images

are then combined with sedimentological analysis

of channel fill deposits to provide some brief

insights into depositional process. The general char-

acter of the studied outcrop has previously been

described by Ricci Lucchi (1969, 1975, 1981) and

Mutti et al. (2002), however, only now can we

examine detailed bed geometries and architecture

by digiscoping. The results will highlight the appli-

cability of the digiscoping technique to outcrop-

based fieldwork, especially in areas where key

exposures are inaccessible.

Fig. 1. Location map of the study area near Fontanelice in the

2. The Fontanelice Channels

The Fontanelice Channels are exposed in an im-

pressive 2D cliff section adjacent to the Santerno

River, near the town of Fontanelice in the Italian

Apennines (Fig. 1). The channels occur within the

uppermost Marnoso Arenacea Formation, which is

Miocene in age (Tortonian). The sediment source for

flows passing through the channels was the southern

Alps and material was ultimately deposited in a basin

environment, although poor outcrop quality has pro-

hibited correlation with the time-equivalent down-

stream sections (Ricci Lucchi, 1969, 1975, 1981;

Mutti et al., 2002).

Two offset-stacked channels are visible in the cliff

section on the northwest side of the Santerno Valley

(Figs. 2 and 3), incising into adjacent and underlying

parallel-bedded sandstone-rich sheets (note that the

lower channel base and underlying beds have recently

been covered by vegetation and talus, see Ricci Luc-

chi (1981) for an older image displaying this section).

Mutti et al. (2002) interpreted the underlying sand-

stone sheets as lobe deposits forming part of the

Castel del Rio dmixedT system. The orientation of

the outcrop relative to palaeoslope (NNW–SSE) sug-

gests that the cliff section is roughly perpendicular to

Italian Apennines. Modified from Ricci Lucchi (1981).

Fig. 2. Digital photo showing the main studied outcrop in the Santerno Valley. Channel bases are shown by dashed lines. Note that the base of

Channel A is no longer visible due to erosion/vegetation cover but is roughly reproduced here (white dashed line) after study of an older photo

in Ricci Lucchi (1981). Large white rectangle shows location of Fig. 3. Small black rectangles show location of Figs. 5 and 6. F=minor fault.

R.B. Wynn et al. / Sedimentary Geology 179 (2005) 153–162 155

channel axes (Mutti et al., 2002). The lower channel is

several hundred metres wide and about 80 m deep,

and is dramatically reincised by the upper channel

which is of similar dimensions (Figs. 2 and 3). For

the purposes of this study, the lower channel will

hereafter be referred to as Channel A and the upper

channel as Channel B. Both channels contain an

aggradational fill comprising thick massive sand-

stones that are commonly amalgamated. Muddy over-

bank/slope deposits are preserved above Channel B,

and are incised by a younger channel complex (Ricci

Lucchi, 1981; Mutti et al., 2002).

Mutti et al. (2002) considered the location and

morphology of the Fontanelice Channels to be struc-

turally controlled, and suggested that the erosional

base of Channel B may represent a slump scar that

was generated in response to increased slope insta-

bility during tectonic uplift. An equally viable hy-

pothesis would be that the channel was initially cut

by high-energy bypassing flows. The channel fill

itself was interpreted to represent deposition from

sand-bearing gravity flows, possibly related to remo-

bilisation of unconsolidated sand from upslope

(Mutti et al., 2002).

3. Digiscoping equipment and technique

Digiscoping combines the improved resolution and

adaptability of modern digital cameras with the high

magnification of field telescopes. The digital camera

is attached to the telescope allowing the photographer

to take ultra-high magnification images of distant

subjects. This method is now widely used in the

field of wildlife photography, but can be applied to

any situation where the subject is a long way from the

observer. The technique is ideally suited to photogra-

phy of distant inaccessible outcrops which, until now,

have been too distant for conventional single-lens

reflex (SLR) photography with zoom lenses.

Fig. 3. Digital photo (a) and line drawing interpretation (b) showing the erosional base and channel fill of Channel B. For location see Fig. 2.

Black lines denote thin-bedded heterolithic intervals. Note the terraced erosional profile of the channel, the tabular nature of thick sandstone

beds and the extensive draping of thin-bedded heterolithic intervals up the channel margin.


Fig. 4. The digiscoping set-up, with a digital camera connected to a tripod-mounted telescope by an aluminium sleeve. This outfit is relatively

cheap and portable, and is a rapid and effective tool for achieving high-quality images of distant outcrops. Features b10 cm in size can be

resolved at distances of several hundred metres.


At the present time, the most suitable cameras for

digiscoping are in the Nikon Coolpix range as these

models have a rotating body and internal zoom lens.

The camera is attached to the telescope using a ma-

chined aluminium bracket and thumb screws (Fig. 4).

With the camera lens set at 4� magnification and a

fixed telescope lens of 20� or 30� magnification

overall magnifications of 50–100� are achievable,

far exceeding the capability of the most powerful

SLR zoom lenses. In practical terms this means that

features b10 cm across can be resolved when the

viewer is at distances of 100 m or more. Individual

images are saved at high-resolution and can then be

arranged into digital photo-mosaics using standard

computer graphics packages such as Adobe Photo-

shop and Illustrator.

4. Digiscoping results and bed geometries

The whole Fontanelice cliff section is shown in

Fig. 2, with the fill of Channel A clearly incised by

Channel B, which is offset-stacked to the southwest.

The base of Channel A is no longer visible, although

an older image published in Ricci Lucchi (1981)

indicates that the channel margin is a low-angle

dterracedT erosion surface. Accessible sections of the

Channel A fill show massive tabular sandstones up to

a few metres in thickness, separated by thin hetero-

lithic intervals. Overall, the sandstone : shale ratio

appears to be ~9 :1.

The base and fill of Channel B (Figs. 2 and 3) is

superbly exposed but completely inaccessible; the

detailed external geometry of features such as large-

scale erosional scours and marginal bed pinch-outs

therefore provide an ideal target for digiscoping.

Consequently, the analysis of external bed geometry

will focus on this channel. The fill of Channel B is

also dominantly composed of massive tabular sand-

stones, with most beds being 1–2 m thick but a

couple reaching 4 m in thickness. These thick sand-

stones are separated by thin heterolithic beds (b1 m

thick) that appear to be fine-grained siltstones and

mudstones. Overall, the sandstone/shale ratio is

~9 :1.

4.1. Bed geometry and marginal pinch-outs

The external geometry (in strike section) of thick

sandstone beds in Channel B is roughly tabular (Figs. 2


and 3). Marginal pinch-outs are commonly abrupt (Fig.

3), although in some cases they are complicated by the

pre-existing dterracedT erosion profile of the channel

base. For example, a digital photo-mosaic illustrates

the pinch out of a thick sandstone bed at the channel

margin (Fig. 5); this pinch out is irregular due to the bed

onlapping a step in the channel erosion profile.

Figs. 3 and 5 also illustrate how thin-bedded het-

erolithic intervals are draped up the channel margin,

and are only completely eroded out at the channel

base and at pronounced steps along the margin. These

thin-bedded intervals reach their greatest thickness on

flat sections of the terraced channel margin. They are

not continuous with adjacent massive sandstones, but

are interbedded between them.

In summary, the irregular pinch-outs and onlap of

beds comprising the channel fill suggest that the

channel has been subject to multiple phases of cut-

and-fill. The tabular nature of the thick sandstone

beds, and their abrupt marginal pinch-outs, contrasts

with the more extensive draping character of interven-

ing thin-bedded heterolithic intervals.

Fig. 5. Digital photo-mosaic (a) and line drawing interpretation (b) showi

Photo-mosaic is composed of 32 individual digital images. For location see

topography generated during initial channel incision. Thin-bedded heter

individual sandstone bodies in the two channel fills. The only places wher

erosion profile, producing a sand-on-sand contact.

4.2. Scours, amalgamation surfaces and soft sediment

deformation

Deposition of thick sandstone beds in Channel B

was accompanied by local scouring and soft sediment

deformation. Individual erosional scours are up to 1 m

deep and 5 m across (Fig. 6), and commonly cut

through underlying shales to generate amalgamation

surfaces. Many of these surfaces can be picked out by a

line of aligned nodules (Figs. 3 and 6), while in other

places the edge of the amalgamation surface is indicat-

ed by truncation of a thin shale unit (Fig. 6). It is likely

that other, less obvious, amalgamation surfaces are also

present in some of the apparently uniform thick sand-

stone beds. In many cases, thin shales were heavily

deformed and contorted during deposition of overlying

sandstone beds, and have been injected upwards over

distances of a metre or more (Fig. 6).

The resolution limitations of digiscoping mean that

from the vantage position (several hundred metres

south-east of the outcrop) features b5 cm across are

not clearly resolvable. Consequently, subtle water

ng pinch-out of a thick sandstone bed at the margin of Channel B.

Fig. 2. Note the complex bed geometry due to onlap of pre-existing

olithic intervals drape the channel margin and effectively separate

e connectivity is achieved is just below a marked step in the channel

Fig. 6. Digital photo-mosaic (a) and line drawing interpretation (b) showing large erosional scour, bed amalgamation and shale deformation/

injection near the base of Channel B. Photo-mosaic is composed of seven individual digital images. For location see Fig. 2.


escape features and small shale clasts, if present,

cannot be observed.

5. Sedimentology of the channel fills

Examination of the few accessible areas at the base

of the studied outcrop provides some indication of the

sediment fill of Channel A, and helps to verify the

digiscoping observations. The dominant sediment fa-

cies are thick massive sandstones separated by thin

fine-grained sandstone, siltstone and mudstone inter-

vals. Similar findings were presented by Mutti et al.

(2002), and beds with identical sediment facies are

also found in a small road outcrop on the opposite

side of the Santerno Valley, that apparently also

exposes a section of one of the channel fills.

5.1. Thick (N1 m) massive sandstones

Thick massive sandstone beds commonly appear

ungraded throughout, although both normal and inverse

grading occur locally at the base of some beds. The

sandstones are poorly sorted and grey-brown in colour.

They generally show a massive, structureless appear-

ance, although faint planar laminations and coarse sand-

stone dstringersT are observed in some places (Fig. 7).

Bed bases are usually erosional, sometimes down to a

few tens of centimetres, and are commonly loaded into

underlying fine-grained intervals; they can also form

irregular amalgamation surfaces with underlying thick

sandstone beds (Figs. 7 and 8). Bed bases contain

abundant randomly scattered coarser grains of 1–6

mm and rare larger pebbles up to 12 mm. Flute marks

are noted at the base of some beds and are infilled by

coarse lag deposits of 2–3 mm grain size. A key feature

is that bed tops are very sharp with a distinct grain size-

break, but may be disrupted by bioturbation.

The overall lack of grading is supported by both

visible and measured grain size analyses, with the

modal grain size of 100–200 Am remaining constant

throughout sampled intervals. The mud content

(here taken as b20 Am) is also remarkably constant,

ranging between 7% and 12% for all samples.

Fig. 7. Digital photo showing thick massive sandstone beds in road section opposite Fontanelice Channel outcrop (outcrop orientations indicate

that these beds are probably part of the channel fill). Note the presence of bed amalgamation and intervening thin-bedded siltstone and mudstone

intervals.


5.2. Thin (b50 cm) siltstones/fine sandstones and

homogeneous mudstones

The finer-grained intervals are normally graded

and bioturbated, and contain rippled greyish fine

sandstone/siltstone overlain by dark grey homoge-

neous mudstones (Figs. 7 and 8). These deposits

occur immediately on top of massive sandstone

beds, and display sharp flat bases. The siltstones and

fine sandstones always show cross- or contorted lami-

nations with unidirectional ripples. Modal grain size is

50–70 Am. The dark grey homogeneous muds are a

few centimetres thick and have gradational contacts

with the underlying rippled silts and fine sands. Bed

Fig. 8. Digital photo showing base of massive sandstone bed with possible sheared layer and clear evidence for loading and injection into

underlying fine-grained intervals.


tops are sharp but often eroded, loaded and/or injected

by overlying thick sandstone beds (Figs. 6, 7 and 8).

6. Depositional processes of the channel fills

Massive sandstone beds in the Fontanelice Channel

fills show little or no vertical grain-size grading, have

sharp bed tops and display a tabular geometry in strike

section with abrupt marginal pinch-outs; these fea-

tures all point towards deposition from hyperconcen-

trated density flows (Mulder and Alexander, 2001).

However, some beds also show features more consis-

tent with concentrated density flows, such as faint

planar laminations and erosional bases with flute

marks (Mulder and Alexander, 2001). A more detailed

interpretation would require vertical grain size profiles

through a series of massive sandstone beds, but this is

not possible at this location due to the largely inac-

cessible nature of the outcrop. The depositional pro-

cess of the fine-grained intervals overlying most

massive sandstone beds is more straightforward; the

presence of normal grading, unidirectional current

ripples and homogeneous mud caps suggest that

they are deposits of relatively dilute turbidity currents.

The observed sequence of (hyper)concentrated

density flow deposits overlain by turbidity current

deposits is repetitive (Fig. 3), suggesting a genetic

link between the different deposit types. If this is

the case then dilute turbidity currents were presum-

ably linked to the (hyper)concentrated density flows

responsible for depositing the massive sandstones

(Mulder and Alexander, 2001). Dilute turbidity cur-

rents may have formed through fluid entrainment and

mixing at the top of the (hyper)concentrated density

flow during its passage downslope, or through distur-

bance of unconsolidated fine-grained seafloor sedi-

ments during initial slope failure and erosion/

pressure wave disruption at the head of the flow.

The resulting stratified flow was therefore composed

of two layers: 1) a sand-rich (hyper)concentrated den-

sity flow that was confined to the channel floor and

deposited thick massive sandstone beds that pinch-out

abruptly at the channel margin, and 2) an overlying

finer-grained, more dilute turbidity current that depos-

ited thin-bedded rippled siltstones and homogeneous

mudstones that taper gradually and drape several

metres up the channel margin.

7. Conclusions

This study shows that digiscoping is a relatively

cheap and efficient technique for imaging distant and/


or inaccessible outcrops, opening new opportunities

for data collection in previously well visited outcrop

areas. It is especially useful for analysing gross exter-

nal geometries at an individual bed scale, but is also

capable of resolving features b10 cm across at dis-

tances of several hundred metres. The application of

digiscoping to the Fontanelice Channels outcrop,

combined with more traditional sedimentological

techniques, has revealed that the channels were filled

by two-layer stratified flows, with a lower (hyper)-

concentrated density flow overlain by a more dilute

turbidity current. The resulting channel fill deposits

are composed of tabular massive sandstones with

abrupt marginal pinch-outs, overlain by thin-bedded

turbidites that taper gradually and drape up the chan-

nel margin.

Acknowledgements

This study forms part of the UK-TAPS Marnoso

Project. Financial support for this project has been

received through a NERC Ocean Margins LINK

grant (NER/T/S/2000/0106) and industry sponsors

(ConocoPhillips, BHP Billiton and Shell UK). We

are particularly grateful to Juli Ericsson and Geoff

Haddad of ConocoPhillips for their long-standing

support and scientific input to the project. The

reviewers, Finn Surlyk, Franco Ricci Lucchi and

Gareth Keevil, are thanked for providing critical

reviews that greatly improved the initial manuscript.

Maarten Felix provided editorial assistance. Harriet

Wimhurst-Brookes is thanked for undertaking grain

size analysis of selected massive sand beds.

References

Bouma, A.H., Stone, C.G., 2000. Fine-grained turbidite systems.

AAPG Memoir 72 and SEPM Spec. Publ. 68, 342 pp.

Hiscott, R., Pickering, K., Bouma, A., Kneller, B., Postma, G., Soh,

J., 1997. Basin-floor fans in the North Sea: sequence stratigraph-

ic models vs. sedimentary facies: discussion. AAPG Bull. 81,

662–665.

Kneller, B., Branney, M.J., 1995. Sustained high-density turbidity

currents and the deposition of thick massive beds. Sedimentol-

ogy 42, 607–616.

Kneller, B., Buckee, C., 2000. The structure and fluid dynamics of

turbidity currents: a review of some recent studies and their


Lowe, D., 1982. Sediment gravity flows: II. Depositional models


currents. J. Sediment. Petrol. 52, 279–297.

Marr, J.G., Harff, P.A., Shanmugam, G., Parker, G., 2001. Experi-

ments on subaqueous sandy gravity flows: the role of clay and

water content in flow dynamics and depositional structures.

GSA Bull. 113, 1377–1386.

Mulder, T., Alexander, J., 2001. The physical character of subma-

rine sediment density flows and their deposits. Sedimentology

48, 269–301.

Mutti, E., Ricci Lucchi, F., Roveri, M., 2002. Revisiting Turbidites

of the Marnoso Arenacea Formation and their Basin-Margin

Equivalents: Problems with Classic Models. Turbidite Work-

shop Guidebook, 64th EAGE Conference and Exhibition,

Parma, Italy.

Ricci Lucchi, F., 1969. Channelized deposits in the Mid-Miocene

Flysch of Romagna (Italy). G. Geol. 36, 203–282.

Ricci Lucchi, F., 1975. Miocene Palaeogeography and Basin Anal-

ysis in the Periadriatic Apennines. In: Squyres, C. (Ed.), Geol-

ogy of Italy, vol. 2. P.E.S.L, Tripoli, pp. 129–236.

Ricci Lucchi, F., 1981. The Miocene Marnoso Areneacea Turbi-

dites, Romagna and Umbria Apennines. In: Ricci Lucchi, F.

(Ed.), IAS Excursion Guidebook, 2nd European Regional Meet-

ing, pp. 231–303.

Shanmugam, G., 1996. High-density turbidity currents: are they

sandy debris flows? J. Sediment. Res. 66, 2–10.

Shanmugam, G., 2000. 50 years of the turbidite paradigm (1950s–

1990s): deep-water processes and facies models—a critical

perspective. Mar. Pet. Geol. 17, 285–342.

Shanmugam, G., 2002. Ten turbidite myths. Earth-Sci. Rev. 58,

311–341.

Shanmugam, G., Bloch, R., Mitchell, S., Beamish, G., Hodgkinson,

R., Damuth, J., Straume, T., Syvertsen, S., Shields, K., 1995.

Basin-floor fans in the North Sea: sequence stratigraphic models

vs sedimentary facies. AAPG Bull. 79, 477–512.

Shanmugam, G., Bloch, R., Damuth, J., Hodgkinson, R., 1997.

Basin-floor fans in the North Sea: sequence stratigraphic models

vs sedimentary facies: reply. AAPG Bull. 81, 666–672.

Stow, D.A.V., Johansson, M., 2000. Deep-water massive sands:

nature, origin and hydrocarbon implications. Mar. Pet. Geol.

17, 145–174.

=Sedimentary Geology 17
Bed geometry used to test recognition criteria of turbidites

and (sandy) debrites

L.A. Amy a,b,*, P.J. Talling a, J. Peakall b, R.B. Wynn c, R.G. Arzola Thynne a

aCentre for Environmental and Geophysical Flows, Department of Earth Sciences, University of Bristol, Bristol, BS8 1RJ, UKbSchool of Earth Sciences, University of Leeds, Leeds, LS2 9JT, UK

cChallenger Division, Southampton Oceanography Centre, European Way, Southampton, Hampshire, SO14 3ZH, UK


Abstract

The origin of thick-bedded deep-water sandstones has generated much controversy in recent years. Two fundamentally

different models have been proposed for beds with the same internal sedimentary characteristics: (1) progressive particle

settling from the base of a turbulent flow—the bturbidity currentQ model and (2) en-masse freezing of a higher-concentration

flow—the bsandy debris flowQ model. These models predict beds with very different geometries; turbidites thin gradually

whereas debrites have abrupt terminations. Previous studies have relied upon sedimentary recognition criteria (i.e., sedimen-

tary features in small-scale outcrop or core) to interpret depositional mechanism. In this study, depositional mechanism is

deduced from bed geometry gained from extensive correlations of individual sandstones preserved in a classic turbidite

system (Marnoso-arenacea Formation, Italy). This approach allows recognition criteria for turbidites and submarine debrites

to be independently tested. We find that tabular and tapered sandstones (turbidites) have distinctly different internal

characteristics to beds with abrupt margins (debrites). Turbidites are relatively well sorted, often exhibit grading and traction

structures and have relatively low matrix mud contents. They may also contain massive division, floating clasts and inverse

grading. Debrites are moderate-to-poorly sorted, ungraded, structureless, contain floating clasts and have elevated matrix mud

contents. These findings have implications for the assessment of submarine gravity flows deposits and reservoir rock

characterization.


Keywords: Turbidity current; Debris flow; Bed geometry; Sedimentary facies


doi:10.1016/j.sedgeo.2005.04.007

* Corresponding author. Tel.: +44 117 954 5235; fax: +44 117 925

3385.

E-mail address: [email protected] (L.A. Amy).

1. Introduction

Submarine sediment-laden density flows are mix-

tures of sediment and water that flow along the sea

or lake floor due to their excess density. These flow

events represent the major sediment transport pro-

9 (2005) 163–174


cess from continental shelf to deep ocean, and their

deposits form thick sedimentary successions that

now host many of the world’s largest petroleum

reservoirs (Weimer and Link, 1991). The ability to

predict lateral variations in bed geometry and sedi-

mentary character (e.g., porosity and permeability)

using sedimentary models is critical to the economic

recovery of hydrocarbons.

The origin of the basal interval of thick-bedded

deep-water sandstones has generated much controver-

sy in recent years (Shanmugam and Moiola, 1995;

1997; D’Agostino and Jordan, 1997; Bouma et al.,

1997; Coleman, 1997; Lowe, 1997; Slatt et al., 1997).

Two contrasting depositional models have been pro-

posed. The bhigh-density turbidity currentQ model

proposes that particles gradually settle out of turbulent

suspension and progressively aggrade a bed (Kuenen

and Migliorini, 1950; Lowe, 1982; Kneller and Bran-

ney, 1995; Stow and Johansson, 2000). Turbidity

currents must have low enough concentrations for

individual particles to settle out from the flow.

Given high suspended load fallout rates, a relatively

high-concentration flow boundary may develop at the

base of the current with reduced turbulence and en-

hanced grain interaction (Lowe, 1982; Kneller and

Branney, 1995). These bhigh-concentration turbidity

currentsQ are considered by some authors to be a flow

type transitional between turbidity currents and debris

flows (e.g., concentrated density flows of Mulder and

Alexander, 2001). Importantly however, deposition

from such currents is still considered to occur pro-

gressively (Lowe, 1982; Kneller and Branney, 1995).

In contrast, a dsandy debris flowT model infers that

deposition occurs by rapid and en-masse freezing of a

high-concentration current (Shanmugam and Moiola,

1995, 1997; Shanmugam, 1996; Stow and Johansson,

2000). Debris flows undergo flow arrest and freezing

when the forces of shear resistance, viscosity and

friction become equal to the flow’s driving force

(Lowe, 1982; Postma, 1986; Mulder and Alexander,

2001). This type of behaviour occurs both in cohe-

sionless flows that have frictional strength due to

interlocking of grains, and cohesive flows that possess

a yield strength due to the presence of cohesive

sediment such as clays.

Distinguishing between a dturbidity currentT or a

ddebris flowT depositional model for sandstone beds is

important since the two models predict deposits with

markedly different geometries. Turbidity currents

should produce spatially extensive deposits with ta-

pered margins as shown by physical experiments

(e.g., Alexander and Morris, 1994; Hallworth and

Huppert, 1998). In comparison, debris flows should

form elongate deposits with abrupt lateral pinch-outs

as indicated by the deposits of subaerial (Major, 1997)

and subaqueous experimental flows (e.g., Hallworth

and Huppert, 1998; Mohrig et al., 1999; Marr et al.,

2001) and the deposits of natural subaerial (e.g., Law-

son, 1982) and very thick (10–50 m) submarine debris

flows (e.g., Aksu and Hiscott, 1989; Laberg and

Vorren, 1995). Some debris flows may deposit by

the progressive accretion of successive flow surges,

but their deposits still display a debrite-like morpho-

logy with abrupt margins (Major, 1997).

Bed geometry is a good indicator of depositional

mechanisms. However, bed geometry is often difficult

to obtain due to the limitations of outcrop or seismic

resolution. Instead, interpretation of sedimentary beds

is usually based on relatively small, centimetre- to

metre-scale sedimentary features preserved in outcrop

or core. Mud-poor, relatively well-sorted sandstones

that contain traction structures and normal grading are

usually interpreted as the deposits of turbidity currents

(e.g., Bouma, 1962; Middleton and Hampton, 1976;

Lowe, 1982; Mutti, 1992). Sedimentary features com-

mon to debrites include (1) mud-rich matrix (if cohe-

sive), (2) poor sorting, (3) shear fabric at the base and

margins, (4) clasts protruding above the top of the bed,

(5) clasts floating in a matrix, (6) a lack of internal

structures, (7) fluid escape structures, (8) sharp upper

contact and (9) lack of grading or inverse grading

throughout or at the base of the deposit (Fisher, 1971;

Rodine and Johnson, 1976; Enos, 1977; Naylor, 1980;

Lowe, 1982; Major, 1997; Sohn, 2000; Marr et al.,

2001; Mulder and Alexander, 2001). However, many

of these features (especially characteristics 5–9) also

occur in metre-thick sandstone beds interpreted as

dturbiditesT. Consequently, it has been suggested that

many well-known dturbiditeT successions such as the

Jackfork Group in Arkansas and Oklahoma (Shanmu-

gam and Moiola, 1995, 1997), the Marnoso-arenacea

Formation in northern Italy (Shanmugam, 1997) and

the Annot sandstones Formation in southeast France

(Shanmugam, 2002) contain large proportions of

sandy debris flow deposits. This reinterpretation is

controversial since it includes structureless, moderate-

60

6970

67

45

66

6559

64

63

1, 80-82

3

2

77

85

84

83

75

10

76

7826

71

72&73

74

42

40

68

6

7

89

11

12

14

79

51&49

15

48

1316

17

5530

50, 52-5428

29 37

56

39

3634

33323120

21

22 23

24

62

44

43

61

57

58

25

Bagno diRomagna

Sansepolcro

Firenzuola

Modigliana

Faenza

Forli

Cesena

10km0

Main flowdirection

N

Measured section

Transect shown in Fig. 2

Main thrust fault

Extent of MA Fm.

MarnosoArenacea Fm.

KEY

BOLOGNA

FIRENZE

Emilia-Romagna

Marche

AdriaticSea

Toscana

Umbria

ITALY

MarnosoArenacea Fm.

Fig. 1. Location map of the study area in central–northern Italy showing the Marnoso-arenacea Formation (shaded area), the positions of

measured sections and line of the stratigraphic panel shown in Fig. 2.



ly to well-sorted sandstones with low mud contents

previously interpreted as the Ta division in the Bouma

(1962) depositional model.

In this study, we use bed geometry as an indepen-

dent criterion for verifying recognition criteria of

turbidites and submarine debrites. The aims of this

study are (1) to distinguish turbidites and sandy deb-

rites in outcrop using bed geometry and (2) to verify

which sediment characteristics present in outcrop at

single locations are diagnostic of depositional flow

type.

2. Study area

The Miocene Marnoso-arenacea Formation crops

out extensively over central–northern Italy (Fig. 1).

The evaluation of bed shape is possible in this forma-

tion since distinctive limestone marker beds and out-

crop continuity allow correlations of individual event

beds over an area of 123 by 27 km (Ricci Lucchi,

1978; Ricci Lucchi and Valmori, 1980). These are the

most extensive correlations of individual beds within

any outcropping turbidite system known to the

authors. The formation mainly comprises basin plain

deposits, composed of interbedded deep-water sand-

stones and mudstones separated by hemipelagic

marls. They were deposited in the Apennine foredeep

basin and subsequently accreted into the Apennine

thrust belt (Ricci Lucchi, 1978; Ricci Lucchi and

Valmori, 1980). The Apennine foredeep basin con-

sisted of a northwest–southeast oriented elongate

trough over 150 km long and 50 km wide (Argnani

and Ricci Lucchi, 2001). Sediment was transported

into the basin by southward-directed flows carrying

detritus from a non-carbonate Alpine source posi-

tioned some 150–200 km northwest of the present

outcrops (Argnani and Ricci Lucchi, 2001). North-

ward directed flows carrying calcareous-rich detritus

were responsible for forming limestone marker beds

(Gandolfi et al., 1983).

A single 25–30 m thick stratigraphic interval of

Serravallian age, located between the most prominent

dContessaT marker bed and a stratigraphically higher

dColombineT marker bed (Ricci Lucchi and Valmori,

1980) was selected to study bed geometry. Recent

publication of 1:10,000 scale geological maps (pub-

lished by the geological surveys of Emilia-Romagna,

Toscana, Marche and Umbria), indicating the position

of the Contessa marker bed, has allowed over 70 new

sections to be measured (Fig. 1). These new sections

have allowed bed shape to be defined with much

greater precision than in previous studies.

During the time of deposition of the studied inter-

val, the basin plain area is believed to have had low

sea-floor gradients a character also exhibited by mod-

ern, deep-sea, abyssal plains (Ricci Lucchi, 1978;

Ricci Lucchi and Valmori, 1980). Low sea-floor gra-

dients are inferred from the high continuity of indi-

vidual beds, absence of channelisation and ability of

currents from different sources to flow in opposite

directions. Hence, in the basin plain area, bathymetry

is not believed to have significantly influenced the

behaviour of sediment gravity flows nor the geome-

tries of the beds they deposited. Hence, this area is an

excellent natural laboratory to study the deposits of

sediment gravity currents that flowed, unobstructed,

across a basin plain.

3. Bed geometry

In this contribution, we concentrate on the geo-

metry of relatively thick sandstone beds. The corre-

lated stratigraphic interval contains 12, metre-thick

sandstone beds of which 10 occur in the stratigraphic

panel shown in Fig. 2. All of these sandstone beds

thin in a downstream direction between the most

northern and southern sections. Two cross-sectional

sandstone bed geometries are common: those that thin

gradually (tapered) and those that thin abruptly down-

stream (Fig. 2). Detailed examples of these beds are

shown in Fig. 3.

3.1. Tapered sandstones

Most of the metre-thick sandstone beds preserved

in relatively proximal sections display a gradual thin-

ning pattern (Figs. 2 and 3A). These beds thin from

N0.8 m to b0.5 m over ~20–40 km. Small changes of

b10–20% of their thickness are recorded between

neighbouring measured sections, spaced 1–10 km

apart. An isopach map for the sandstone thickness

of bed 7 is shown in Fig. 4A and illustrates the

localised thickening and thinning often seen in these

beds. Localised thickness changes may be related to

Thr

ustf

ault

Thr

ustf

ault

Contessa mudstone Contessa mudstone

Flow directionfor most beds

KEY

1

Datum

2

2.5

3

4

5

5.1

6

7

8

C1

m si vf f m c

Sect. No.

Bed No.

17.0 km 16.9 km21.1 km

25m

7.2 km 6.0 km

8 11 12 14 48 50 3755 60 70 455762 44 61 58

?

?

SENW

Marl

Mudstone

Poorly sorted silty-muddy sandstonecontaining mud clasts

Cover

Mudstone or marl (undistinguished)

Sandstone (well-to-moderately sorted)

Fig. 2. Stratigraphic panel showing correlation of thick beds between relatively proximal (left) and relatively distal (right) sections over a distance of some 68 km. The datum from

which sections are hung is the top of the Contessa marker bed’s mudstone. Measured sections are simplified to show only lithofacies and not internal bed structure. Flow direction of

depositional currents was from left to right except for carbonate-rich beds (e.g., bed C1) whose parental flow travelled in the opposite direction. Most thick sandstone beds display a

gradual reduction in thickness moving downstream. Beds 2.5 and 5.1, however, display an abrupt decrease in thickness downstream.

L.A.Amyet

al./Sedimentary

Geology179(2005)163–174

167

(B) BED 2.5: Sandstone showing abrupt thinning

Thin sand or mud ~10 cm thick

8 11 12 14 48 50 3755 60 455762 44 61 58

(A) BED 4: Sandstone showing gradual thinning

Lithologic units

Marl

Mudstone

Mudstone or marl (undistinguished)

Sandstone (well-to-moderately sorted)

Poorly sorted silty-muddy sandstone

Cover

Liquefaction lamination

Chaotic swirly texture

Folded / sheared?mud-poor sandstone

Sedimentary structures

Cross-lamination

Dunecross-bedding

Starved dune / dune top Planar stratified

Planar laminated

Overturned convolute laminated

Convolute laminated

Wavy:convolute or dune laminated

Clast horizon

Diagenetic nodules orsiltstone clasts

Scattered clastsmud / marl

?

FlatUndularLoadedSmall flutes (<2cm deep)Grooved / Tool marks

Base of bedClasts

Organics

Wavy surfaces / indistinct surfacesdune-scale wavy, dewatering or,corrugated related structures

Flow directionm si vf f m c

17.0 km

(Note transect broken by thrust fault)(Note transectbroken bythrust fault)

16.9 km21.1 km

1m

1m

7.2 km 6.0 km

KEY

Fig. 3. Sedimentary logs of two correlated sandstone beds. Bed 4 is an example of a bed that displays gradual downstream thinning (A). Bed 2.5 is an example of a bed that shows

abrupt thinning downstream (B). These two bed types are interpreted as the deposits of turbidity currents and cogenetic turbidity current–debris flows, respectively.

L.A.Amyet

al./Sedimentary

Geology179(2005)163–174

168

Fig. 4. Isopach maps of sandstone thickness of a turbidite bed (bed 7) displaying gradual thickness changes (A) and a cogenetic debrite-turbidite

bed (bed 2.5) displaying abrupt thickness changes (B). Sandstone thickness is defined as the portion of the bed with a grain size of siltstone and

coarser. Isopach maps were created using ODMk computer software. Grids use an inverse distance interpolation method. Interpolation does not

use data from different sides of major thrust faults.



the compensation of subtle basin floor irregularities

caused by the depositional relief of previous deposits

or local basin subsidence. More significant changes in

bed thickness, ~40–50% of their thickness, are occa-

sionally observed. For example, bed 7 decreases by

~50% of its thickness between sections 57 and 58

over a distance of ~3 km, and so does bed 1 between

70 and 45 over a similar distance (Fig. 2). However,

such large changes are usually associated with the

gain or loss of the very fine sandstone or siltstone

found at the top of the bed and not the coarser grain

size fraction. The large spatial extent and gradual

downstream thinning of these beds suggests deposi-

tion by turbidity currents.

3.2. Sandstones with abrupt margins

Four out of the twelve correlated metre-thick sand-

stone beds display an abrupt change in their thickness

moving downstream. The abrupt margin of two of

these units, beds 2.5 and 5.1, occur in the stratigraphic

panel shown in Fig. 2. Bed 2.5 is shown in detail in

Fig. 3B. These sandstones thin by ~80% from ~1 m

to b0.2 m thickness, over 2–10 km. In some cases, the

sandstone bed pinches-out completely and only a

correlative mud cap may be identified. In plan view,

the thick portion of the bed is tongue-shaped (e.g.,

Fig. 4B), thinning in a direction both parallel and

perpendicular to the flow direction. The elongate

plan-form shape, limited spatial extent and abrupt

terminations of the thick portion of these sandstone

beds suggest deposition from a debris flow.

4. Internal character

4.1. Turbidite sandstones (beds with tapered margins)

The internal sedimentary character of the turbidite

sandstones varies between individual beds and local-

ities but a number of similar features are shared (Fig.

5A). They are relatively well sorted, contain traction

structures (cross, parallel and convolute lamination)

and commonly show normal grading. Beds commonly

display 1–10-cm-thick planar or gently undulating

grain size stratification in their lower part. In some

beds, the basal stratified unit is inversely graded and is

observed to occur in a step-wise fashion, as opposed

to a gradual grain size increase. However, traction

structures and normal grading are not always perva-

sive throughout the bed; beds commonly exhibit mas-

sive intervals lacking structure or grading (Fig. 3A).

Clasts are sometimes present and may occur individ-

ually or as a group scattered along a discrete horizon.

The total mud content of collected samples from

several turbidite beds was measured using SEM anal-

ysis (Talling et al., 2004). In this analysis, detrital mud

was not distinguished from diagenetic mud and hence

probably indicate somewhat greater values than that

derived from the depositional flow. Results show that

turbidites have a relatively low mud matrix content of

b12% and typically between 5–8% by volume.

4.2. Debrite-turbidite sandstones (beds with abrupt

margins)

Beds with abrupt margins display distinctive ver-

tical and lateral changes in character (Fig. 3B). The

thick portion of the beds in most locations is charac-

terized by a distinctive tripartite structure comprising

(1) a basal, massive or laminated unit, (2) a middle,

unstructured and relatively thick, clast-rich unit and

(3) a laminated upper unit (Fig. 5B). This tripartite

bed structure occurs upstream of where the sandstone

begins to pinch and is continuous for 20–30 km (Figs.

3B and 4B). In downstream sections, the sandstone

bed is much thinner (b20 cm) and it is characterized

by fine-grained laminated sand. In relatively proximal

northern sections, the tripartite bed structure is some-

times absent and instead the bed comprises structure-

less, normally graded or ungraded sandstone.

The middle interval in the tripartite bed structure,

previously referred to as a dslurried divisionT (RicciLucchi and Valmori, 1980), is distinctive in character.

It is moderate to relatively poorly sorted (containing

mud, silt sand and clasts) and ungraded with a rela-

tively sharp grain size discontinuity at its upper

surface. Outsize clasts of mud and hemipelagite,

0.2–30 cm long, are randomly distributed throughout

the unit and larger sand grains may be dispersed

within the finer grained matrix. The matrix has a

relatively high mud content, N15% and typically

20–22% by volume as assessed from SEM analysis

of samples (Talling et al., 2004). The high mud con-

tent ensures that outcrops are often friable and have a

distinct grey hue.

Fig. 5. Outcrop photographs and sedimentary logs of a turbidite bed (A) and a cogenetic turbidite-debrite bed (B). The turbidite bed has a

characteristic blocky weathering pattern. The debrite has a distinctive swirly weathering pattern and occurs sandwiched between thinner

cogenetic turbidite sandstones. See Fig. 3 for graphic log key.


The abrupt termination of the middle clast-rich

interval of tripartite beds suggests it is a debrite. The

basal and upper divisions of tripartite beds, however,

have nearly identical characteristics to those found in

tapered turbidite sandstones, such as traction struc-

tures and a low mud content (b10% by volume).

Similarly, the characteristics of relatively proximal

parts of these beds (i.e., low mud contents and normal

grading) suggest they may be turbidites and not deb-

rites although this interpretation cannot be justified


using geometry. The abrupt lateral margins of these

composite ddebrite-turbiditeT beds reflect the pinch-

out of the clast-rich debrite interval; hence, a turbidite

interpretation for proximal and basal clean sandstones

within these beds does not compromise the geometric

data.

These tripartite beds indicate that individual sub-

marine events traversing distal basin plains may have

both turbidity current and debris flow components. A

detailed discussion of the origins of cogenetic turbi-

dite-debrite beds from the Marnoso-arenacea Forma-

tion is provided by Talling et al. (2004). Beds with a

similar tripartite structure are noted from other deep-

water systems and have also been interpreted as

cogenetic turbidite-debrite beds (e.g., Hickson,

1999; Kneller and McCaffrey, 1999; McCaffrey

and Kneller, 2001; Haughton et al., 2003; Talling

et al., 2004).

5. Discussion

5.1. Implications for recognition criteria

Certain sedimentary characteristics preserved in

outcrops are indicative of depositional process, as

determined independently using bed geometry. Turbi-

dite sandstones may be recognised on the basis of

being relatively well sorted and possessing relatively

low mud contents, normal grading and traction struc-

tures. Massive ungraded sandstones that are relatively

well sorted with low mud contents (often described as

the Bouma Ta division in other studies) also common-

ly occur within turbidite sandstones. In addition, tur-

bidite sandstones may contain floating clasts and

inverse grading locally at their base usually associated

with centimetre-thick stratification. The results of this

study suggest that all of these sedimentary features

can be formed by particles progressively falling out of

suspension from a relatively dilute and turbulent flow,

albeit in some instances through a relatively high-

concentration basal flow boundary. These findings

are consistent with the recognition criteria developed

over the last four decades for the deposits of sand-

bearing turbidity currents (e.g., Bouma, 1962; Mid-

dleton and Hampton, 1976; Lowe, 1982; Mutti,

1992). The results do not support the model that

relatively well-sorted and mud-poor massive sand-

stones and those that exhibit inverse grading or float-

ing clasts within this system are deposited exclusively

by (sandy) debris flows as suggested by Shanmugam

and Moiola (1995, 1997) and Shanmugam (1996,

1997, 2000, 2002). Field data indicate that debrites

are characterized by moderate-to-poorly sorted, un-

graded and structureless sandstones with floating

clasts and a relatively high matrix mud content, con-

curring with many previous models (e.g., Middleton

and Hampton, 1976; Lowe, 1982; Mutti, 1992).

Experimental studies have suggested that very

low mud contents are sufficient to suspend sand-

sized particles and cause debris flow behaviour.

Hampton (1975) calculated that sediment mixtures

of fine sand with clay contents as low as 2 wt.%, or

less, can move as debris flows. Marr et al. (2001)

produced sandy debris flows with mud contents of

0.7 wt.% using bentonite and 7 wt.% using kaolin-

ite. However, much higher mud contents were

found in turbidite beds of the Marnoso-arenacea

Formation; mud contents of up to 12% by volume

are recorded in turbidites, although measured values

include both detrital and diagenetic mud. Hence, the

interpretation of ancient sandstones as debrites

based on low absolute values of mud content, i.e.,

in the range of a few percent as suggested by

experiments, appears problematic if not simply be-

cause it may be difficult to distinguish detrital from

diagenetic mud. However, it was found that debrites

of the Marnoso-arenacea Formation did have rela-

tively high mud percentages (N20% by volume)

compared to cleaner turbidite sandstones. Hence,

relative values of mud content, as opposed to ab-

solute values are possibly a better indicator of

depositional mechanism.

The bed correlations show that submarine flows

are capable of depositing both debrite and turbidite in

the same event so that different deposit types may

pass laterally into one another over short distances

(Fig. 3B). Hence, recognition criteria may be used to

identify depositional mechanism locally but may not

always be valid away from points of control. For

example, a thin laminated turbidite sandstone may

represent the lateral equivalent of either a relatively

thick turbidite sandstone or debris flow unit. Future

work should be aimed at distinguishing those sedi-

mentary characteristics that can be used to predict the

character of beds laterally.


Acknowledgment

The research was funded by the United Kingdom

Natural Environmental Research Council and Conoco

(now ConocoPhillips) through the OCEAN MAR-

GINS-LINK scheme (grant number NER/T/S/2000/

01403). We thank Luca Martelli and Giani Zuffa for

advice concerning the field area and George Postma,

Guido Ghibaudo, Stan Stanbrook and Jaco Baas for

constructive reviews.

References

Aksu, A.E., Hiscott, R.N., 1989. Slides and debris flows on the

high-latitude continental slopes of Baffin-bay. Geology 17,

885–888.

Alexander, J., Morris, S., 1994. Observations on experimental non-

channelized, high concentration turbidity current and variations

in deposits around obstacles. Journal of Sedimentary Research

A64, 899–909.

Argnani, A., Ricci Lucchi, F., 2001. Tertiary siliciclastic turbidite

systems of the Northern Apennines. In: Vai, G.B., Martini, P.

(Eds.), Apennines and Adjacent Mediterranean Basins. Kluwer

Academic Publishers, pp. 327–350.

Bouma, A.H., 1962. Sedimentology of some flysch deposits: a

graphic approach to facies interpretation. PhD thesis, Elsevier,

Amsterdam, 168 pp.

Bouma, A.H., Devries, M.B., Stone, C.G., 1997. Reinterpretation of

depositional processes in a classic flysch sequence (Pennsylva-

nian Jackfork Group), Ouachita Mountains, Arkansas and Okla-

homa: Discussion1. AAPG Bulletin 81, 470–472.

Coleman, J.I., 1997. Reinterpretation of depositional processes in a

classic flysch sequence (Pennsylvanian Jackfork Group), Oua-

chita Mountains, Arkansas and Oklahoma: Discussion1. AAPG

Bulletin 81, 466–469.

D’Agostino, A.E., Jordan, D.W., 1997. Reinterpretation of deposi-

tional processes in a classic flysch sequence (Pennsylvanian

Jackfork Group), Ouachita Mountains, Arkansas and Oklahoma:

Discussion1. AAPG Bulletin 81, 473–475.

Enos, P., 1977. Flow regimes in debris flows. Sedimentology 24,

133–142.

Fisher, R.V., 1971. Features of coarse-grained, high concentration

fluids and their deposits. Journal of Sedimentary Petrology 41,

916–927.

Gandolfi, G., Paganelli, L., Zuffa, G.G., 1983. Petrology and dis-

persal directions in the Marnoso Arenacea Formation (Miocene

Northern Apennines). Journal of Sedimentary Petrology 53,

493–507.


concentration, particle-driven gravity currents. Physics of Fluids

10, 1083–1087.

Hampton, M.A., 1975. Competence of fine-grained debris flows.

Journal of Sedimentary Petrology 45, 834–844.

Haughton, P.D.W., Barker, S.P., McCaffrey, W.D., 2003. Linked

debrites in sand-turbidites systems—origin and significance.


Hickson, T.A., 1999. A Study of Deep-water Deposition: Con-

straints on the Sedimentation Mechanics of Slurry Flows and

High Concentration Turbidity Currents, and the Facies Archi-

tecture of a Conglomeratic Channel Overbank System. Stanford

University, California. (256 pp.)

Kneller, B.C., Branney, M.J., 1995. Sustained high density turbidity


ogy 42, 607–616.

Kneller,, McCaffrey, W.D., 1999. Depositional effects of flow non-

uniformity and stratification within turbidity currents approach-

ing a bounding slope. Journal of Sedimentary Research 69,

980–991.

Kuenen, H., Migliorini, C., 1950. Turbidity currents as a cause of

graded bedding. The Journal of Geology 58, 91–127.

Laberg, J.S., Vorren, T.O., 1995. Late weichselian submarine debris

flow deposits on the Bear-island-trough-mouth-fan. Marine Ge-

ology 127, 45–72.

Lawson, D.E., 1982. Mobilization, movement and deposition of

active subaerial sediment flows, Matanuska Glacier, Alaska.

Journal of Geology 90, 279–300.

Lowe, D.R., 1997. Reinterpretation of depositional processes in a

classic flysch sequence (Pennsylvanian Jackfork Group), Oua-

chita Mountains, Arkansas and Oklahoma: Discussion1. AAPG

Bulletin 81, 460–465.

Lowe, D.R., 1982. Sediment gravity flows: II. Depositional models



Major, J.J., 1997. Depositional processes in large-scale debris flow

experiments. Journal of Geology 105, 345–366.

Marr, J.G., Harff, P.A., Shanmugam, G., Parker, G., 2001.

Experiments on subaqueous sandy gravity flows: the role of

clay and water content in the flow dynamics and depositional

structures. Geological Society of America Bulletin 113,

1377–1386.

McCaffrey, W.D., Kneller, B.C., 2001. Process controls on the

development of stratigraphic trap potential on the margins of

confined turbidite systems and aids to reservoir evaluation.

AAPG Bulletin 85, 971–988.

Middleton, G.V., Hampton, M.A., 1976. Subaqueous sediment

transport and deposition. In: Stanley, D.J., Swift, P.J.P. (Eds.),

Marine Sediment Transport and Environmental Management.

John Wiley and Sons, New York, pp. 197–218.

Mohrig, D., Elverhøi, A., Parker, G., 1999. Experiments on the

relative mobility of muddy subaqueous and subaerial debris

flows, and their capacity to remobilize antecedent deposits.

Marine Geology 154, 117–129.

Mulder, T., Alexander, J., 2001. The physical character of sub-

marine density flows and their deposits. Sedimentology 48,

269–301.

Mutti, E., 1992. Turbidite Sandstones. Instituto di Geologia, Uni-

versita di Parma, AGIP.

Naylor, M.A., 1980. The origin of inverse grading in muddy flow

deposits—a review. Journal of Sedimentary Petrology 50,

1111–1116.


Postma, G., 1986. Classification of sediment gravity-flow deposits

based on flow conditions during sedimentation. Geology 14,

291–294.

Ricci Lucchi, F., Valmori, E., 1980. Basin-wide turbidites in a

Miocene, over-supplied deep-sea plain: a geometrical analysis.


Ricci Lucchi, F., 1978. Turbidite dispersal in a Miocene deep-sea

plain: the Marnoso Arenacea of the Northern Apennines. Geo-

logie en Mijnbouw 57, 559–576.

Rodine, J.D., Johnson, A.M., 1976. The ability of debris, heavily

freighted with coarse materials, to flow on gentle slopes. Sed-

imentology 23, 213–234.

Shanmugam, G., 1996. High-density turbidity currents; are they

sandy debris flows? Journal of Sedimentary Research A66,

2–10.

Shanmugam, G., 1997. The Bouma sequence and the turbidite mind

set. Earth-Science Reviews 42, 201–229.

Shanmugam, G., 2000. 50 years of the turbidite paradigm (1950s–

1990s): deep-water processes and facies models—a critical

perspective. Marine and Petroleum Geology 17, 285–342.

Shanmugam, G., 2002. Ten turbidite myths. Earth-Science Reviews

58, 311–341.

Shanmugam, G., Moiola, R.J., 1995. Reinterpretation of deposition-

al processes in a classic flysch sequence (Pennsylvanian Jack-

fork Group), Ouachita Mountains, Arkansas and Oklahoma.

AAPG Bulletin 79, 672–695.

Shanmugam, G., Moiola, R.J., 1997. Reinterpretation of deposition-

al processes in a classic flysch sequence (Pennsylvanian Jack-

fork Group), Ouachita Mountains, Arkansas and Oklahoma:

Reply1. AAPG Bulletin 81, 476–491.

Slatt, R.M., Weimer, P., Stone, C.G., 1997. Reinterpretation of

depositional processes in a classic flysch sequence (Pennsylva-

nian Jackfork Group), Ouachita Mountains, Arkansas and Okla-

homa: Discussion1. AAPG Bulletin 81, 449–459.

Sohn, Y.K., 2000. Depositional processes of submarine debris flows

in the Miocene fan deltas, Pohang Basin, SE Korea with special

reference to flow transformation. Journal of Sedimentary Re-

search 70, 491–503.

Stow, D.V., Johansson, M., 2000. Deep-water massive sands: na-

ture, origin and hydrocarbon implications. Marine and Petro-

leum Geology 17, 145–174.

Talling, P.J., Amy, L.A., Wynn, R.B., Peakall, J., Robinson, M.,

2004. Beds comprising debrite sandwiched within co-genetic

turbidite: origin and widespread occurrence in distal deposition-

al environments. Sedimentology 51, 163–194.

Weimer, P., Link, M.H., 1991. Global Petroleum occurrences in

submarine fans and turbidite systems. In: Weimer, P., Link,

M.H. (Eds.), Seismic Facies and Sedimentary Process of Sub-

marine Fans and Turbidite Systems. Springer-Verlag, New York,

pp. 9–67.

Documents

Sedimentary Geology 179