Unconfined turbidity current interactions with oblique slopes: deflection,reflection and combined-flow behaviours

Download the full PDF ➡

Read the full paper ⬇

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/380677872

Unconfined turbidity current interactions with oblique slopes: deflection,

reflection and combined-flow behaviours

Preprint · May 2024

DOI: 10.31223/X5569F

CITATION

READS

6 authors, including:

Ru Wang

University of Leeds

8 PUBLICATIONS 83 CITATIONS

SEE PROFILE

Edward Keavney

University of Leeds

13 PUBLICATIONS 17 CITATIONS

SEE PROFILE

All content following this page was uploaded by Ru Wang on 14 January 2025.

The user has requested enhancement of the downloaded file.This is a non-peer reviewed preprint submitted to EarthArxiv

TITLE

Unconfined turbidity current interactions with oblique slopes: deflection, reflection and

combined-flow behaviours

AUTHORS AND AFFILIATIONS

Ru Wang1*

, Jeff Peakall1

, David M. Hodgson1

, Ed Keavney1

, Helena C. Brown1 and Gareth

M. Keevil1

1 School of Earth & Environment, University of Leeds, Leeds, LS2 9JT, UK

* Corresponding author. Ru Wang: earrwa@leeds.ac.uk

Submitted to Sedimentology for peer-review, 16th May 2024

22This is a non-peer reviewed preprint submitted to EarthArxiv

ABSTRACT

What is the nature of flow reflection, deflection and combined-flow behaviour when gravity

flows interact with slopes? In turn, how do these flow dynamics control sedimentation on

slopes? Here, these questions are addressed using physical experiments, with low-density

unconfined gravity flows interacting with slopes of varying gradients, at a range of flow

incidence angles. The present paradigm for gravity current interaction with slopes was based

on experiments with high-density flows, conducted in narrow 2D flume tanks, in small (1 m2

planform) 3D tanks, or in large 3D tanks where flows can surmount the topography. Here,

larger-scale physical experiments were undertaken in unconfined settings where the flow

cannot surmount a planar topographic slope. The experiments show that the dominant flow-

process transitions from divergence-dominated, through reflection-dominated to deflection-

dominated as the flow incidence angle varies from 90° to 15° and the slope gradient changes

from 20° to 40°

. Also, patterns of velocity pulsing at the base of, and on, the slope vary as a

function of both the flow incidence angle and slope gradient. Furthermore, in all configurations

complex multidirectional combined flows are observed on, or at the base of, the slope, and are

shown to vary spatially across the slope. The findings challenge the paradigm of flow deflection

and reflection in existing flow-topography process models that has stood for three decades. A

new process model for flow-slope interactions is presented, that provides new mechanics for

the frequent observation of palaeocurrents from sole marks at high angles to those in the

associated ripple division. Results provide insights into the formation and spatial distribution

of distinctive combined-flow bedforms, sediment dispersal patterns, and process controls on

onlap termination styles in deep-sea settings, which can be applied to refine interpretations of

exhumed successions.

Keywords: unconfined turbidity current, topographic slope, incidence angle, slope gradient,

flow deflection, flow reflection, combined flow, velocity pulsing

47This is a non-peer reviewed preprint submitted to EarthArxiv

INTRODUCTION

Turbidity currents are subaqueous gravity-driven turbulent flows that serve as important

mechanisms for the transfer of large volumes of clastic sediments from the continental shelf to

the deep oceans (e.g., Kuenen and Migliorini, 1950; Dzulynski et al., 1959; Sestini, 1970;

Normark et al., 1993; Kneller and Buckee, 2000). Seafloor topography influences turbidity

current behaviour, and therefore the distribution and nature of their deposits. The interplay of

several factors need to be considered in the interaction of turbidity currents and topography

(Tinterri, 2011; Patacci et al., 2015; Tinterri et al., 2022 and references therein), including flow

duration (surge versus sustained or quasi-steady flow), the relative volume of the flow versus

the size of the basin (‘flow confinement’, hereafter; sensu Tőkés and Patacci, 2018; cf.

Pickering and Hiscott, 1985; Southern et al., 2015), and the configuration of the containing

topography (e.g., slope gradient, orientation and geometry; ‘topographic containment’,

hereafter). When the volume of the flow is small relative to the size of the basin, the flow can

expand in the basin freely, which is referred to as unconfined flow in this work. In the presence

of seafloor topography, flows can be reflected, deflected and/or constricted depending on the

configuration of the containing topography and the flow properties (e.g., thickness, viscosity,

and velocity).

A better understanding of the complicated interactions between turbidity currents and seafloor

topography, and the links to depositional character, is critical in a wide range of situations. For

example, palaeogeographic reconstruction of ancient deep-water basins (e.g., Sinclair, 1994;

Lomas and Joseph, 2004; Bell et al., 2018), hydrocarbon or CO2 reservoir characterisation in

the subsurface (e.g., McCaffrey and Kneller, 2001; Chadwick et al., 2004; Bakke et al., 2013;

Lloyd et al., 2021), modern mass-flow geohazard assessment in deep-water environments (e.g.,

Bruschi et al., 2006; Carter et al., 2014), prediction of plastic litter and other pollutant

72This is a non-peer reviewed preprint submitted to EarthArxiv

distribution in the deep sea (e.g., Haward et al., 2018; Kane et al., 2020) and de-risking

management of sedimentation in modern human-made water reservoirs (e.g., Wei et al., 2013).

The opaque nature of natural turbidity currents and limited field instrumental measurements

have restricted the understanding on the interaction between turbidity currents and containing

topography. Advances have been made mainly through scaled-down physical experiments

(e.g., Pantin and Leeder, 1987; Muck and Underwood, 1990; Alexander and Morris, 1994;

Kneller et al., 1991; Edwards et al., 1994; Amy et al., 2004; Patacci et al., 2015; Soutter et al.,

2021), numerical modelling (e.g., Athmer et al., 2010; Howlett et al., 2019) and facies analysis

of exhumed systems (e.g., Kneller et al., 1991; Haughton 2000; Tinterri et al., 2016, 2022).

The previous experimental studies have been conducted either in narrow 2D flume tanks (e.g.,

Edwards et al., 1994; Amy et al., 2004; Patacci et al., 2015), in small (1 m2 planform) 3D tanks

(Kneller et al., 1991; Kneller, 1995), or in large 3D tanks with low-relief topographic

configurations that are surmounted by the flows (Soutter et al., 2021). Field outcrop-based

models of confined and contained turbidites are derived from purely theoretical analysis with

limited 3D constraints (e.g., Kneller and McCaffrey, 1999; Hodgson and Haughton, 2004), or

from linking to existing 2D flume experimental data (e.g., Tinterri et al., 2016, 2022).

Therefore, their significance in understanding the temporal and spatial variability in the

dynamics of flow-topography interactions is limited. Hence, the behaviour of 3D unconfined

turbidity currents that interact with different configurations of topographic slopes has not been

investigated.

Combined flows and the formation of hummock-like or sigmoidal bedforms in deep-water

systems have previously been linked to the interaction of turbidity currents with topography

and the superposition of a unidirectional parental turbidity current with an oscillatory

component due to the reflections of the internal waves or bores against a topographic slope

(Kneller et al., 1991; Edwards et al., 1994; Patacci et al., 2015; Tinterri, 2011; Tinterri et al.,

97This is a non-peer reviewed preprint submitted to EarthArxiv

2016, 2022), largely based on the observations from 2D or qualitative 3D reflected density

current experiments (e.g., Kneller et al., 1991; Edwards et al., 1994). Based on experimental

observations of 3D, unconfined density currents interacting with an orthogonal planar slope,

100

Keavney et al. (2024) propose a new mechanism for the generation of combined flows on

101

slopes, with the absence of internal waves. However, whether the new mechanism holds in

102

cases where 3D, unconfined density currents interact with an oblique topographic slope has not

103

been investigated experimentally.

104

In this work, a series of Froude-scaled 3D physical experiments were conducted using

105

sustained, unconfined saline density currents, where the flow was partially contained by a rigid

106

planar slope. The flows did not overtop the barrier but were able to flow downstream around

107

the slope. Here, dissolved salt acts as a surrogate for fine mud in suspension that does not easily

108

settle out, and moves in bypass mode, and therefore flows used in this work can be considered

109

to model low-density turbidity currents (Sequeiros et al., 2010). The overall aim of this work

110

is to systematically investigate the effects of different configurations of topographic slopes on

111

the flow behaviour, including incidence angle of the flow onto the slope and slope gradient. To

112

achieve this, the following three objectives are undertaken: (i) to investigate the influence of

113

containing topography on the general flow behaviour, including flow decoupling and stripping,

114

lateral flow expansion on the slope surface, and the relative strength between flow deflection

115

versus reflection; (ii) to explore the effect of containing topography on the temporal near-bed

116

velocity pulsation patterns, a property that is crucial for sediment erosion and deposition; and

117

(iii) to assess the effect of containing topography on the temporal variability of near-bed flow

118

directions.

119

The results are subsequently discussed considering their implications for the development of

120

new models of flow-topography interactions, and the generation and spatial distribution of

121

complex, multidirectional combined flows in deep-water settings. Finally, these findings are

122This is a non-peer reviewed preprint submitted to EarthArxiv

used to provide insights into the formation and spatial distribution of distinctive combined-

123

flow bedforms, such as hummock-like and sigmoidal bedforms, sediment dispersal patterns,

124

and process controls on onlap termination styles, which can be applied to the interpretation of

125

exhumed successions in deep-sea settings.

126

127

METHODOLOGY

128

Experimental design and data collection

129

Experiments were conducted in the Sorby Environmental Fluid Dynamics Laboratory,

130

University of Leeds. The flume tank used is 10 m long, 2.5 m wide and 1 m deep, with a flat

131

basin floor (Fig. 1A). A 1.8 m long straight input channel section was centred in the upstream

132

end of the main tank, through which the saline density currents entered the tank. The first

133

experiment was run without any basin-floor topography (unconfined experiment) and served

134

as a base-case experiment for scaling. Eighteen subsequent ramp experiments were conducted

135

with a non-erodible, smooth, planar ramp (1.5 m wide and 1.2 m long) placed on the base of

136

the flume tank. The ramp had a tapered leading edge at the foot abutting the basin floor, which

137

minimized any step discontinuity. The leading edge at the foot of the ramp was placed 3 m

138

downstream from the channel mouth (black dashed line in Fig. 1A), with its centrepoint located

139

on the channel-basin centreline (red circle in Fig. 1A). This position was chosen as the density

140

current had lost the effects of upstream confinement and was relatively unconfined (see

141

Turbidity current evolution in the unconfined experiment subsection). In these ramp

142

experiments, the slope gradient (S) and incidence angle (IN) were systematically varied. Each

143

experiment (Table 1) considers a different combination of incidence angle relative to the

144

incoming flow (i.e., 90°, 75°, 60°, 45°, 30° and 15°; Fig. 1B) and ramp slope gradient (i.e., 20°,

145

30° and 40°; Fig. 1C-E). The maximum barrier height in these topographic configurations is

146This is a non-peer reviewed preprint submitted to EarthArxiv

0.410 m, 0.585 m, and 0.76 m, respectively, and was tested to be able to fully contain the flow

147

vertically, i.e., the density current did not surmount the topography.

148

149

Fig. 1. (A) Schematic sketch of the experimental facility. Note that the base of the containing

150

topographic ramp is indicated as a black dashed line. Position of the Ultrasonic Velocity

151

Profiler (UVP), Acoustic Doppler Velocimeter (ADV) and siphoning system for the unconfined

152

experiment is also indicated. (B-E) Topographic configurations of the ramp experiments with

153

different combinations of slope gradients and incidence angles relative to the incoming flow.

154This is a non-peer reviewed preprint submitted to EarthArxiv

(B) Ramp with different incidence angles relative to the incoming flow shown in a plan view.

155

The left-side and right-side of the tank are relative to the incoming flow. (C-E) Ramp with

156

different slope gradients shown in a side view. Measuring localities of the four ADVs (ADVs

157

1-4) for each ramp experiment are illustrated. Two sets of Cartesian coordinate systems are

158

adopted: relative to the basin floor (A) or to the ramp (F).

159

160

Before each experiment, the main tank was filled with fresh tap water to 0.6 m deep. A saline

161

solution of excess density 2.5% (1025 kg m-3) was prepared in a 2 m3 mixing tank with an

162

electric rotary mixer utilised to ensure a uniform salt concentration. The saline density current

163

was subsequently pumped into the main tank at a constant discharge rate of 3.6 L s-1 (Table 1).

164

Water density and temperature were measured using a portable densimeter (DMA35, Anton

165

Parr, Graz, Austria; a resolution of 0.1 kg m-3 and 0.1 °C, respectively) in both the main tank

166

and the mixing tank before each experimental run (Table 1). The discharge rate was controlled

167

by an inverter-governed centrifugal pump and monitored in real time by an electromagnetic

168

flowmeter (Fig. 1A). The density current entered the main tank through a diffuser pipe, and

169

then flowed through the straight channel. The diffuser prevented development of a jet flow

170

being directed straight down the tank. Each experiment started with the release of the flow

171

from the mixing tank to the main tank and ended after a total run time of 130 s.

172

Unconfined experiment

173

In the unconfined experiment, four repeats were run using near identical initial conditions but

174

for different purposes (Fig. 1A): i) flow visualisation with an overhead camera; ii) velocity

175

profiling using an ultrasonic velocity profiler (UVP); iii) velocity profiling using an acoustic

176

Doppler velocity profiler (ADV); and iv) density profiling using a siphon array. In the flow

177

visualisation run, overhead images were taken by a Fujifilm X-T4 camera with Fujifilm 14 mm

178

f/2.8R XF lens to capture the whole view of the experiment every second. Fluorescent purple

179This is a non-peer reviewed preprint submitted to EarthArxiv

TABLE 1. Experimental parameters. Tinflow water temperature in mixing tank. Tmaintank water temperature in main tank. Note that four repeats were

180

conducted for the unconfined experiment and three repeats for each ramp experiment, respectively, due to experimental constraints.

181

Experiment

Slope

angle

(°)

Incidence

angle (°) Data collected Mean flow rate

(L s-1) Tinflow (℃) Tmaintank (℃) Inlet flow density (kg m-3)

Unconfined N/A N/A

Flow visualisation; a

UVP, ADV & density

siphoning system

positioned at 3 m

downstream from the

channel mouth along the

channel-basin centreline

3.61, 3.60, 3.60,

3.60 13.2, 7.5, 12.9, 6.0 13.8, 7.9, 13.5, 6.8 1025, 1025, 1025,1025

S20°IN90° 20 90

Flow visualisation; 4

ADVs (one positioned at

the base of the slope

along the channel-basin

centreline and the other

three at the flow front

positions above the slope

surface)

3.60, 3.61, 3.60 9.3, 9.6, 9.8 9.9, 10.0, 9.7 1025.1, 1025, 1024.9

S20°IN75° 20 75 3.59, 3.61, 3.60 20.9, 20.2, 20.0 21, 20.4, 20.7 1025, 1024.6, 1025

S20°IN60° 20 60 3.59, 3.60, 3.59 19.8, 19.4, 19.0 20, 19.6, 19.6 1025, 1024.6, 1024.9

S20°IN45° 20 45 3.59, 3.59, 3.59 18.5, 18.4, 18.4 19.0, 18.7, 18.7 1025.2, 1024.8, 1025

S20°IN30° 20 30 3.59, 3.60, 3.60 18.4, 18.8, 18.5 19.1, 19.0, 19.0 1025, 1025.2, 1024.8

S20°IN15° 20 15 3.60, 3.59, 3.59 18.9, 19.0, 19.2 19.4, 19.4, 19.6 1024.8, 1024.9, 1025

S30°IN90° 30 90 3.59, 3.59, 3.60 7.4, 8.0, 7.9 7.7, 7.8, 8.3 1024.9, 1024.9, 1025

S30°IN75° 30 75 3.60, 3.59, 3.59 19.2, 18.9, 19.9 19.5, 19.2, 20.1 1025.4, 1024.5, 1024.5

S30°IN60° 30 60 3.60, 3.60, 3.60 19.8, 19.8, 20.8 20.2, 21.1, 21.1 1025.2, 1024.8, 1025

S30°IN45° 30 45 3.59, 3.60, 3.59 20.1, 20.1, 20.2 20.8, 20.8, 20.6 1025, 1024.8, 1024.5

S30°IN30° 30 30 3.60, 3.60, 3.60 20.0, 19.4, 19.6 20.4, 19.8, 20.0 1024.9, 1025, 1024.6

S30°IN15° 30 15 3.59, 3.59, 3.60 20.0, 19.8, 19.8 20.4, 20.2, 20.1 1024.7, 1025, 1024.9

S40°IN90° 40 90 3.58, 3.59, 3.59 9.6, 9.7, 9.8 10.1, 10.0 10.2 1025, 1024.9, 1025

S40°IN75° 40 75 3.60, 3.60, 3.62 19.4, 19.1, 19.3 19.8, 19.4, 19.6 1024.3, 1025.3, 1025.3

S40°IN60° 40 60 3.60, 3.60, 3.60 19.9, 19.6, 19.7 20.0, 20.0, 20.1 1024.9, 1025.3, 1025.3

S40°IN45° 40 45 3.59, 3.60, 3.59 16.9, 16.9, 16.7 17.2, 17.0, 17.0 1024.9, 1025, 1025

S40°IN30° 40 30 3.59, 3.59, 3.60 18.8, 17.8, 17.8 19.1, 18.1, 18.2 1024.9, 1025.3, 1025

S40°IN15° 40 15 3.60, 3.59, 3.60 18.7, 18.7, 17.8 19.0, 19.1, 18.2 1025.3, 1025, 1025

182This is a non-peer reviewed preprint submitted to EarthArxiv

dye was added to the input density current to aid flow visualisation. To monitor the real-time

183

flow properties (velocity and density) and provide a reference for the subsequent ramp

184

experiments, velocity profiles collected by UVP and ADV systems and density profiles by a

185

siphon system were obtained for flows at 3 m downstream from the channel mouth along the

186

channel-basin centreline (i.e., the position of the base of the ramp in subsequent experiments;

187

Fig. 1A).

188

UVP (Met-Flow, UVP DUO, 4 MHz; Met-Flow SA, Lausanne, Switzerland; Fig. 2A) was

189

utilised to record the velocity field of the entire density current (cf. Takeda, 1991, 1993; Best

190

et al., 2001; Lusseyran et al., 2003; Keevil et al., 2006). A vertical array of 10 UVP probes was

191

oriented parallel to the basin floor and positioned at 0.01, 0.02, 0.03, 0.04, 0.05, 0.06, 0.07,

192

0.09, 0.11 and 0.13 m respectively above the basin floor (Fig. 2A). Each UVP probe recorded

193

the instantaneous downstream flow velocity at 128 measurement positions along its axis

194

extending 10 to 29 cm from the probe head in the configuration used (see Table S1 for details

195

of the UVP set-up).

196

197

Fig. 2. Set up of (A) the UVP, (B) ADV and (C) siphoning systems in this study to measure the

198

velocity and density profiles, respectively. All profiles were measured vertical to the basin

199This is a non-peer reviewed preprint submitted to EarthArxiv

floor, irrespective of whether the instrument was mounted above the basin floor or the slope

200

surface.

201

202

ADV (Nortek Vectrino Profiler; Nortek Inc., Rud, Norway; Fig. 2B) was used to capture the

203

temporal evolution of the 3D velocities of the flows at a near-bed region (i.e., a coverage of

204

0.03 m height above the basin floor or slope surface). ADV records 3-components of velocity

205

in bins with a vertical resolution of 1 mm (see Table S1 for the details of the ADV set-up). The

206

ADV data constrain the 3D velocity structure of the flows through 100 Hz measurements of

207

instantaneous velocities (cf. 4 Hz for the UVP; Table S1). The measurements of the near-bed

208

velocity are critical to understanding the conditions that effect sediment transport and

209

deposition. Therefore, ADV was utilised in the subsequent ramp experiments to capture the

210

near-bed velocity field of the saline density currents. During the experimental runs for the

211

velocity profiling collection, a mixture of neutrally buoyant hollow glass spheres (Sphericel

212

110-P8; 10 µm diameter) were seeded into the inlet flow at a constant discharge rate via a

213

peristaltic pump throughout the experimental run (cf. Thomas et al., 2017; Ho et al., 2019).

214

This was undertaken to enhance the reflection of the ultrasound or acoustic signal. Additionally,

215

prior to each run, the ambient water in front of the UVP or ADV probes was also seeded with

216

the same glass spheres to increase the signal-to-noise ratio to ca. 30 dB.

217

The fluid flow samples were collected by a siphoning system (Fig. 2C). The siphons were

218

positioned along a vertical line and located at 0.005, 0.015, 0.020, 0.029, 0.038, 0.047, 0.055,

219

0.063, 0.070, 0.077, 0.085 and 0.094 m respectively above the basin floor. During the

220

experimental run, the fluid flow was extracted from the tank via a peristaltic pump at a constant

221

flow rate (3.9 mL s-1 per siphon tube). This specific value was chosen to balance obtaining

222

enough fluid samples whilst minimising perturbations to the in-situ flow structure. After each

223This is a non-peer reviewed preprint submitted to EarthArxiv

run, the density of the collected fluid samples was measured by the aforementioned portable

224

densimeter.

225

226

Ramp experiments

227

In each ramp experimental configuration, three repeats were run using identical initial

228

conditions but with different purposes, i.e., flow visualisation and velocity profiling by ADV

229

systems.

230

In the flow visualisation runs, each experiment was recorded using up to four high-resolution

231

video cameras (GoPro, HERO 10; GoPro, Inc., USA). One was mounted at ca. 2 m downstream

232

from the channel mouth along the channel-basin centreline to capture the front view of the

233

density current encountering the containing topography (i.e., ramp), two along the side of the

234

ramp to capture the side view, and one directly on the top of the ramp surface to capture the

235

top view. No dye was added to the inlet flow as it would provide little information on the

236

internal fluid motion within the current. Instead, Pliolite, a low density and highly reflective

237

polymer, and a small amount of white paint were added to the input current to help visualisation

238

(cf. Edwards et al., 1994). The Pliolite has a subspherical shape, with a mean grain size of 1.5

239

mm and density of 1050 kg m-3

. To improve the visualisation of the density current interacting

240

with the topographic ramp, fluorescent yellow dye was injected via a series of tubes mounted

241

from the rear of the ramp and flush with its surface. These tubes were located at three different

242

elevations and distributed evenly on the ramp surface (i.e., 0.15 m, 0.30 m, and 0.45 m away

243

from the base of the ramp, respectively).

244

In each ramp experimental configuration, four ADVs were utilised to record the 3D flow

245

velocity field at the near-bed region (Fig. 1B-E and Fig. 2B). One was positioned above the

246

basin floor, at 0.02 m upstream from the base of the ramp along the channel-basin centreline

247This is a non-peer reviewed preprint submitted to EarthArxiv

(ADV1) to capture the basal flow reversals. The other three (ADVs 2-4) were placed above the

248

slope surface to capture the temporal evolution of the velocity field near the flow front position

249

(see General flow behaviour subsection). The exact locations of these three ADVs were

250

carefully chosen based on the position of the flow front observed from the flow visualisation

251

videos, which varied across different experiments. The transducers of the ADVs 1-4 were

252

mounted vertically 0.07 m above the slope surface and recorded the velocity profile in thirty-

253

one 1-mm-high cells ranging from 0 to 0.03 m above the slope surface (Fig. 2B). Due to

254

experimental constraints, two sets of ADV data (ADVs 1-2 and ADVs 3-4) were collected in

255

separate runs with the same initial conditions, varying the measurement locations of the ADVs

256

in each case. The 4 ADVs were subsequently integrated to visualize the velocity field of the

257

whole flow. During each measurement, synchronization of the two ADVs was achieved using

258

Nortek’s MIDAS data acquisition software (Nortek 2015) and the recording started from the

259

release of the inlet flow until the flow ceased.

260

261

Experimental data analysis

262

All the raw instantaneous velocity data collected by the UVP and ADV systems were initially

263

filtered in Matlab before further analysis (cf. Buckee et al., 2001; Keevil et al., 2006). First,

264

data spikes in the time series that were more than two standard deviations from the mean were

265

removed; here, the mean was estimated as an 11-point moving average. Second, the removed

266

spike points were replaced by a 3-point moving mean. The ADV data closest to the boundary

267

were affected by excess noise because of reflections. Consequently, the plotted data were

268

clipped so that the bottom 5 data points (< 0.5 cm) were removed (Fig. 2B). This excess noise

269

sometimes affected points as high as 0.7 cm above the bed, and thus for data analysis only the

270

section between 0.7-3.0 cm above the basin floor or slope surface were utilised.

271This is a non-peer reviewed preprint submitted to EarthArxiv

In this work, two sets of Cartesian coordinate systems were adopted, either relative to the basin

272

floor or to the ramp (Fig. 1A and 1F). The filtered 3D velocity data after the first step were

273

corrected based on either of these two coordinate systems. When the former coordinate system

274

is adopted, the 3D velocity components (𝑢, 𝑣, 𝑤) are termed as streamwise, cross-stream and

275

vertical velocities, respectively. Otherwise, they are termed as down-dip, along-strike, and

276

vertical velocities, respectively.

277

The filtered instantaneous velocity data collected by the ADV system are presented as velocity

278

time-series profiles. In these plots, positive values of the down-dip velocity depict flows

279

travelling towards the ramp (outbound flow), whereas negative ones depict flows travelling

280

away from the ramp and back towards the inlet (return flow). The maximum velocity (Umax)

281

up/down the ramp, is taken as the highest value over the measured height range (0.7-3.0 cm)

282

of the ADV profiles. The fluctuations in Umax are shown on the time series panels and serve as

283

a representative flow down-dip velocity magnitude.

284

285

Flow scaling and characterisation

286

As only saline density currents are utilised in this work, Froude scaling (Yalin, 1971; Peakall

287

et al., 1996) is used to ensure that both the dimensionless Froude and Reynolds numbers of the

288

laboratory turbidity currents reside within appropriate flow regimes compared to natural

289

systems (in the Froude scaling approach, the Froude number in the experimental flows should

290

be similar to that of natural systems, while the Reynolds number is relaxed). When these scaling

291

conditions are met, the laboratory turbidity currents can be considered scalable to natural

292

systems.

293

The Reynolds number, Re, is used to characterize whether the flow is laminar or turbulent and

294

is expressed by the ratio between the inertial forces to the viscous forces. It is given by

295This is a non-peer reviewed preprint submitted to EarthArxiv

𝑅𝑒 =

𝜌𝑠𝑈ℎ

𝜇 (1)

296

where 𝜌𝑠 represents the depth-averaged density of the current, 𝑈 is the depth-averaged velocity

297

over the flow height, ℎ is the flow height, and 𝜇 is dynamic viscosity. Typically, flows with Re

298

> 2000 are considered fully turbulent, flows with Re < 500 are laminar, and flows with Re =

299

500-2000 are transitional.

300

The Froude number, Fr, describes the ratio between inertial- and gravitational-forces, and is

301

expressed as

302

𝐹𝑟 =

𝑈

√𝑔ℎ (2)

303

where 𝑔 denotes gravitational acceleration. Typically, flows with Fr > 1 are considered

304

supercritical whereas flows with Fr < 1 are subcritical, though this critical value might be

305

different in strongly stratified density currents (e.g., Sumner et al., 2013; Cartigny et al., 2014).

306

For experiments involving density difference, such as turbidity currents, the densimetric

307

Froude number is more physically relevant, defined by

308

𝐹𝑟𝑑 =

𝑈

√𝑔′ℎ (3)

309

𝑔′

𝑔(𝜌𝑠−𝜌𝑎)

𝜌𝑎

(4)

310

where 𝑔′ represents the reduced gravitational acceleration and 𝜌𝑎 denotes the density of the

311

ambient fluid.

312

Based on the unconfined control experiment, the experimental density currents recorded at 3

313

m downstream from the channel mouth along the channel-basin centreline (i.e., the position

314

where the centrepoint of the base of the slope resides; Fig. 1A) were demonstrated to have a

315

Reynolds number of 3203 and densimetric Froude number of 0.505 (Table 2), and therefore

316This is a non-peer reviewed preprint submitted to EarthArxiv

were fully turbulent and subcritical. Estimation of these two parameters is detailed in

317

Supporting Information 1.

318

319

TABLE 2. Summary of the flow characteristics for the experimental density current recorded

320

at 3 m downstream from the channel mouth along the channel-basin centreline in the

321

unconfined reference experiment. Calculations of the mean depth-averaged downstream

322

velocity and current density are detailed in Supporting Information 1.

323

Parameter Value Unit

Density of the ambient fluid (𝜌𝑎) 999.58 kg m-3

Dynamic viscosity (𝜇) 0.001 Pa s

Gravitational acceleration (𝑔) 9.81 m s-2

Reduced gravitational acceleration (𝑔′) 0.030 m s-2

Flow depth (ℎ) 0.11 m

Mean depth-averaged density of the current (𝜌𝑠) 1002.6 kg m-3

Mean depth-averaged downstream velocity (𝑈) 0.029 m s-1

Maximum downstream velocity (𝑢𝑝) 0.059 m s-1

Height of the maximum downstream velocity above the

basin floor (ℎ𝑝)

0.02 m

Reynolds number (𝑅𝑒) 3203 none

Densimetric Froude number (𝐹𝑟𝑑) 0.505 none

324

RESULTS

325

Turbidity current evolution in the unconfined experiment

326

In the unconfined experiment, the saline density current enters the confined channel section as

327

a highly turbulent flow with a well-developed head region, which is followed by a stable, quasi-

328

steady body region during the rest of the experimental run (Fig. 3A). On exiting the confined

329

channel section, the flow starts to spread radially and symmetrically above the flat basin floor

330

(Fig. 3B). Multiple lobes and clefts can be observed at the propagating head of the density

331

currents. A radial hydraulic jump can be observed immediately downstream of the channel-

332

mouth location (Fig. 3D), suggesting that the flow regime has transitioned from a supercritical

333This is a non-peer reviewed preprint submitted to EarthArxiv

state in the channel section to a subcritical state in the horizontal basin floor (see also Flow

334

scaling and characterisation subsection). Finally, the termination of the inlet leads to a rapid

335

decrease in current velocity and causes the current body to diminish quickly.

336

The representative time-averaged UVP downstream velocity profile obtained from the body

337

region of the flows (averaging over 30 s; Fig. 3G) was recorded at 3 m downstream from the

338

channel mouth along the channel-basin centreline. The velocity profile reveals a mean depth-

339

averaged downstream velocity of 0.029 m s-1

, a mean depth-averaged current density of 1002.6

340

kg m-3 (i.e., 0.3% excess density) and a flow height or thickness of ca. 0.11 m (Table 2;

341

Supporting Information 1). The downstream velocity reaches its maximum value (up = 0.059

342

m s-1) at a height of 0.02 m above the basin floor (hp = 0.02 m).

343

The time-averaged flow density profile at the same position (Fig. 3G) exhibits a noticeable

344

exponential decrease in excess density upward, with a highest flow density (ρsi = 1009 kg m-3;

345

0.9% excess density) near the basin floor (hi = 0.005 m). The density currents at 3 m

346

downstream from the channel mouth along the basin centreline are demonstrated to be density-

347

stratified (cf. Stacey and Bowen, 1988) throughout the experimental run: the density time-

348

series plot for the flow current at this position (Fig. 3H) exhibits a distinct dense region near

349

the basal part of the flow and a dilute region at the upper part of the flow.

350

351This is a non-peer reviewed preprint submitted to EarthArxiv

352

Fig. 3. (A-F) Set of overhead photographs illustrating the evolution of the saline density

353

currents from the channel section to the basin floor in the unconfined reference experiment.

354

Note that a radial hydraulic jump was observed immediately downstream of the channel mouth.

355

(G) Profiles of time-averaged flow downstream velocity and density for the experimental

356

density current recorded at 3 m downstream of the channel mouth along the channel-basin

357

centreline in the unconfined reference experiment. Both measurements were initiated 5 s after

358This is a non-peer reviewed preprint submitted to EarthArxiv

the current head passed and lasted for 30 s. The flow depth ℎ, maximum downstream velocity

359

𝑢𝑝 , its height above the basin floor ℎ𝑝 , depth-averaged downstream velocity 𝑈 and depth-

360

averaged density 𝜌𝑠 are shown in the panel as red squares. The ambient water density was

361

measured at 12°C. (H) Time-series profiles of flow density measured at 3 m downstream of the

362

channel mouth along the channel-basin centreline, the position of which is shown as a red circle

363

in Figure 1A.

364

365

Interaction of turbidity currents with containing topography in the ramp experiments

366

General flow behaviour

367

Here, experimental observations for Experiment S20°IN75° (Fig. 4) are described in detail to

368

summarize the general flow behaviour when flows encounter the topographic slope. Once the

369

flow exits the channel, it propagates along the basin as an unconfined underflow until

370

encountering the containing slope (Fig. 4A). Upon incidence with the topographic slope, the

371

flow decelerates and becomes strongly multidirectional on the slope surface (Fig. 4B).

372

Simultaneously, flow stratification promotes the original flow to be decoupled into two parts:

373

a lower denser part, and an upper less dense part. The dilute upper part of the flow runs up the

374

slope surface and thins until reaching its maximum height Hmax (‘maximum run-up height’,

375

hereafter; cf. Pantin and Leeder, 1987; Edwards et al., 1994; Fig. 4C). This is termed as flow

376

thinning and stripping on the slope surface hereafter. In contrast, the dense, lower part of the

377

flow collapses back down the slope and is either deflected parallel to the slope and/or reflected

378

towards the inlet at the base of the slope (Fig. 4C). The zone of flow stripping on the slope

379

surface can be quantified by the height of the initial reversal of the dense lower flow Hmin and

380

the maximum run-up height Hmax. Specifically, the lower limit of the flow stripping zone is

381

quantified by the height upslope at which the basal region of the flow reverses downslope

382

because this marks the onset of flow thinning upslope. The initial reversal of the dense lower

383This is a non-peer reviewed preprint submitted to EarthArxiv

384

Fig. 4. Representative side-view photographs depicting the temporal evolution of density

385

currents upon incidence with an oblique topographic slope (Experiment S20°IN75° for

386

example). Hmax denotes the maximum height that the dilute, upper part of the flow can run up

387

on the slope surface. t denotes the experimental time since the release of the flow from the

388

mixing tank.

389

390

part of the flow can undercut the primary outbound flow and migrate upstream from the slope

391

before eventually dissipating in the basin. This initial flow reversal of the basal part of the flow

392

just above the containing slope leads to a thickening of the entire body of the density current

393

(Fig. 4D), which is termed as an unsteady ‘inflation’ phase of the suspension cloud by Patacci

394This is a non-peer reviewed preprint submitted to EarthArxiv

et al. (2015). Subsequently, as the parental flow re-establishes, the suspension cloud in the

395

basin becomes flat-topped (i.e., a sharp, subhorizontal interface with the ambient water) and a

396

quasi-stable flow front develops on the slope surface (Fig. 4F). This is termed a quasi-steady

397

phase by Patacci et al. (2015). Finally, the waning of the inlet flow causes the suspension cloud

398

to collapse. Note that no trains of upstream-migrating solitons or bores are observed throughout

399

the experiments (cf. Pantin and Leeder, 1987; Edwards et al., 1994). Flow behaviour, including

400

the degree of lateral flow expansion on the slope surface, the degree of flow thinning and

401

stripping, and the relative strength between flow deflection and reflection, varies as a function

402

of both the slope gradient and the incidence angle of the flow onto the slope.

403

404

Variation of incidence angles of the current onto the slope

405

The effects of containing slope orientation, with respect to flow direction, on flow behaviour

406

were explored by systematically changing the incidence angles of the flow to the slope with

407

the same slope gradient. Here, the results for 3 of the 18 experiments are presented: S40°IN75°,

408

S40°IN60° and S40°IN15° (Videos 1-3).

409

In Experiment S40°IN75° (Video 1), upon encountering the topographic slope, the flow runs

410

into the slope strongly and results in a wide divergence in flow velocity directions on the slope

411

surface. The area of lateral flow expansion on the slope surface is the largest among the three

412

experiments. The maximum run-up height (Hmax = 0.29 m) occurs in the middle of the ramp,

413

whereas the height of initial flow reversal develops at ca. 0.13 m. Due to the high degree of

414

topographic containment generated by the oblique ramp orientation in this experiment,

415

reflection of the dense, basal part of the current is the strongest among these three experiments.

416

Part of the dense, basal part of the flow is deflected and runs parallel to the slope. This basal

417

flow is diverted at the point of incidence to the slope into two directions towards the lateral

418This is a non-peer reviewed preprint submitted to EarthArxiv

edges of the slope, with the dividing streamline or plane (cf. Kneller and McCaffrey, 1999) at

419

ca. 0.56 m from the right edge of the ramp.

420

421

Video 1. Annotated video illustrating the behaviour of density currents upon incidence with an

422

oblique topographic slope (Experiment S40°IN75°).

423

424

In Experiment S40°IN60° (Video 2), relatively less flow is observed to be able to run up the

425

slope and more of the flow is deflected towards the lateral edge of the slope, compared to

426

Experiment S40°IN75° . The divergence in flow velocity directions on the slope surface is also

427

less pronounced. The area of lateral flow expansion on the slope surface decreases markedly.

428

Hmax develops at the right edge of the ramp, at ca. 0.24 m upslope; the height of initial flow

429

reversal is 0.13 m upslope. Flow reflection at the basal part of the slope is less pronounced due

430

to a decrease in the topographic containment (see also Temporal velocity pulsing subsection).

431

Hence, basal flow deflection is stronger relative to flow reflection, in contrast to Experiment

432

S40°IN75° . The dividing streamline of the deflected dense, basal region of the flow is ca. 0.37

433

m from the right edge of the ramp.

434This is a non-peer reviewed preprint submitted to EarthArxiv

435

Video 2. Annotated video illustrating the behaviour of density currents upon incidence with an

436

oblique topographic slope (S40°IN60°).

437

438

In Experiment S40°IN15° (Video 3), the highly oblique ramp orientation results in the current

439

mainly being deflected parallel to the base of the slope with extremely limited interaction

440

between the current and slope surface (i.e., limited flow reflection or lateral flow expansion).

441

The zone of flow thinning and stripping on the slope surface is negligible, with the height of

442

initial flow reversal located at 0.12 m upslope and maximum run-up height at 0.16 m upslope.

443

444

Video 3. Annotated video illustrating the behaviour of density currents upon incidence with an

445

oblique topographic slope (Experiment S40°IN15°).

446

447This is a non-peer reviewed preprint submitted to EarthArxiv

Variation of slope gradients

448

The effects of slope gradient on flow behaviour were investigated using a single oblique

449

incidence angle. Here, the results for 3 of the 18 ramp experiments are presented: S20°IN75°,

450

S30°IN75° and S40°IN75° (Fig. 4, Videos 1 and 4).

451

452

Video 4. Annotated video illustrating the behaviour of density currents upon incidence with an

453

oblique topographic slope (Experiment S30°IN75°).

454

455

Results in Experiment S40°IN75° were described in the preceding section. In Experiment

456

S30°IN75° (Video 4), upon encountering the containing slope, the flow strikes the slope less

457

strongly and becomes multidirectional on the slope surface but with a much larger area of

458

lateral flow expansion, compared to Experiment S40°IN75°. Hmax occurs laterally at ca. 0.37 m

459

away from the right edge of the ramp, and ca. 0.36 m upslope; the height of initial flow reversal

460

is ca. 0.12 m upslope. The strength of the flow reflection is not apparent in the visualisation

461

video. However, the deflection of the dense, basal part of the flow can be identified. The basal

462

flow is deflected into two directions towards the two lateral edges of the slope, respectively,

463

with the dividing streamline ca. 0.56 m from the right edge of the ramp.

464This is a non-peer reviewed preprint submitted to EarthArxiv

In Experiment S20°IN75° (Fig. 4), a much larger area of lateral flow expansion on the slope

465

surface is observed, compared to former experiments. Hmax occurs laterally at ca. 0.37 m away

466

from the right edge of the ramp, and ca. 0.26 m upslope; the height of initial flow reversal is

467

ca. 0.1 m upslope. Like the case in Experiment S30°IN75°, the strength of flow reflection

468

cannot be identified visually, but part of the basal flow is deflected to run parallel to the slope.

469

470

Temporal velocity pulsing

471

From the flow visualisation videos, a series of upstream-migrating velocity reversals in the

472

basal part of the flow can be identified, above the flat basin floor near the base of slope, and on

473

the slope surface (Videos 1-4). Furthermore, the depth-constrained ADV down-dip velocity

474

time-series profiles (Figs 5-8) capture the velocity reversals quantitatively at a point.

475

476This is a non-peer reviewed preprint submitted to EarthArxiv

Fig. 5. Down-dip velocity time series of the density currents recorded at the base of the slope

477

along the channel-basin centreline (ADV1 in Figure 1) for the ramp experiments (i.e.,

478

S20°IN90°, S20°IN75°, S20°IN60°, S20°IN45°, S20°IN30° and S20°IN15°). For

479

visualisation, the data are clipped at z ~0.5 cm due to excess noise, caused by reflections. The

480

temporal evolution of maximum velocity up/down the ramp, Umax, [i.e., the highest value over

481

the measured height range (0.7-3.0 cm) of the ADV profiles] is also shown (blue solid lines).

482

483

484

Fig. 6. Down-dip velocity time series of the density currents recorded at the base of the slope

485

along the channel-basin centreline (ADV1 in Figure 1) for the ramp experiments (i.e.,

486

S20°IN90°, S30°IN90° and S40°IN90°). For visualisation, the data are clipped at z ~0.5 cm

487

due to excess noise, caused by reflections. Positive values of the down-dip velocity depict flows

488This is a non-peer reviewed preprint submitted to EarthArxiv

travelling towards the ramp, whereas negative values depict flows travelling away from the

489

ramp and back towards the inlet. The temporal evolution of maximum velocity up/down the

490

ramp, Umax, [i.e., the highest value over the measured height range (0.7-3.0 cm) of the ADV

491

profiles] is also shown (blue solid lines).

492

493

494

Fig. 7. Down-dip velocity time series of the density currents recorded at the flow front position

495

just above the slope surface (ADV3 in Figure 1) for the ramp experiments (i.e., S20°IN75°,

496

S30°IN75° and S40°IN75°). For visualisation, the data are clipped at z ~0.5 cm due to excess

497

noise, caused by reflections. The temporal evolution of maximum velocity up/down the ramp,

498

Umax, [i.e., the highest value over the measured height range (0.7-3.0 cm) of the ADV profiles]

499

is also shown (blue solid lines).

500This is a non-peer reviewed preprint submitted to EarthArxiv

501

Fig. 8. Down-dip velocity time series of the density currents recorded at the flow front position

502

just above the slope surface (ADV3 in Figure 1) for the ramp experiments (i.e., S20°IN60°,

503

S20°IN45°, S20°IN30° and S20°IN15°). For visualisation, the data are clipped at z ~0.5 cm

504

due to excess noise, caused by reflections. The temporal evolution of maximum velocity

505

up/down the ramp, Umax, [i.e., the highest value over the measured height range (0.7-3.0 cm)

506

of the ADV profiles] is also shown (blue solid lines).

507

508

Base of slope: Reflection and basal flow reversal

509

Down-dip velocity time-series profiles of the flow recorded near the base of slope along the

510

channel-basin centreline (Figs 5-6) exhibit multiple basal flow reversals when the flow

511

encounters the topographic slope. Notably, the first basal flow reversal is of high-velocity and

512

highly turbulent, which is succeeded by a series of weaker basal flow reversals. After the first

513

basal flow reversal diminishes, the second reversal typically re-establishes from an initially

514

very low velocity to a final high velocity. The velocity of each reversal is generally lower than

515

the preceding one. Nevertheless, the magnitude of the velocity, the number of velocity pulses,

516This is a non-peer reviewed preprint submitted to EarthArxiv

and the duration of each pulse are different across the ramp experiments, as a function of both

517

incidence angle and slope gradient.

518

Base of slope: Variation of incidence angles of the current onto the slope

519

Variation of incidence angle as a function of a single slope gradient (20°) is examined for

520

experiments S20°IN90°, S20°IN75° , S20°IN60°, S20°IN45°, S20°IN30° and S20°IN15° (Fig.

521

5). Notably, for lower incidence angles, the magnitude of the maximum down-dip velocity Umax

522

markedly decreases (Umax = 0.06 ~ 0.008 m s-1 for the basal flow reversals in Experiment

523

S20°IN90° and Umax = 0.03 ~ 0.01 m s-1 in Experiment S20°IN15°). Furthermore, the velocity

524

pattern tends to be characterised by more pulses (N = 3 for the basal flow reversals in

525

Experiment S20°IN90° and N > 7 in Experiment S20°IN15°) and shorter time duration of each

526

pulse (T = 8 ~ 12 s for the basal flow reversals in Experiment S20°IN90° and T = 2 ~ 7 s in

527

Experiment S20°IN15°).

528

Base of slope: Variation of slope gradients

529

For cases across different slope gradients, results of the experiments S20°IN90°, S30°IN90°

530

and S40°IN90° are presented (Fig. 6). In Experiment S20°IN90° (Fig. 6A), the first basal flow

531

reversal begins ca. 13 s after the arrival of the first outbound flow and subsequently sustains

532

for ca. 10 s until the re-establishment of the second outbound flow. The maximum magnitude

533

of the first velocity reversal reaches ca. 0.06 m s-1. This is followed by four weaker flow

534

reversals, with time duration of each pulse of 11, 12, 3, and 1.4 s respectively and Umax ranging

535

from 0.005 to 0.026 m s-1. In Experiment S30°IN90° (Fig. 6B), the first basal flow reversal

536

arrives at 9 s after the first outbound flow initially establishes, which then sustains for ca. 8 s

537

with a recorded downdip maximum velocity over height of 0.06 m s-1. This is succeeded by

538

three weaker flow reversals, with time duration of each pulse of 14, 6 and 4 s respectively and

539

Umax ranging from 0.011 to 0.023 m s-1. In Experiment S40°IN90° (Fig. 6C), the first basal

540This is a non-peer reviewed preprint submitted to EarthArxiv

flow reversal starts to develop at 10 s after the arrival of the first outbound flow, which then

541

sustains for ca. 5.5 s with a recorded downdip maximum velocity over height of 0.04 m s-1

542

This is succeeded by seven weaker flow reversals, with time duration of each pulse of 4, 4.4,

543

6, 5, 3, 2 and 3 s respectively and Umax ranging from 0.008 to 0.026 m s-1. For cases across

544

different slope gradients, the magnitude of the maximum velocity shows minimal difference.

545

However, experiments with a higher angle of slope gradient are demonstrated to be dominated

546

by more velocity pulses and shorter time duration of each pulse.

547

In summary, the incidence angle of the current relative to the containing slope exerts a much

548

stronger control on the velocity pulsing pattern of the flow near the base of the slope (e.g., the

549

strength and time duration of each basal flow reversal) than the slope gradient.

550

On the slope: Flow front velocity fluctuation

551

During the quasi-steady phase of each ramp experiment, a quasi-stable flow front develops on

552

the slope surface, which fluctuates over a short distance up slope (Fig. 4F). Fluctuations of the

553

flow front velocity are examined quantitatively via the depth-constrained ADV down-dip

554

velocity time-series profiles positioned at the centreline of the ramp (ADV3 in Figure 1; Figs

555

7-8). Compared to measurements located at the base of the slope, the velocity magnitude of the

556

flow front is lower. The velocity structure, number of velocity pulses, and time duration of each

557

pulse (Figs 7-8) are a function of both the incidence angle of the flow and the slope gradient.

558

For cases with different slope gradients (S20°IN75°, S30°IN75° and S40°IN75°), the

559

magnitude of the maximum down-dip velocity Umax exhibits only small variation, between -

560

0.05 and 0.07 m s-1 (Fig. 7). Experiments with a steeper slope gradient configuration are

561

associated with relatively more velocity pulses and shorter time duration of each pulse albeit

562

the differences are small.

563This is a non-peer reviewed preprint submitted to EarthArxiv

Considering experiments S20°IN75°, S20°IN60°, S20°IN45°, S20°IN30° and S20°IN15°,

564

those with a lower flow incidence angle tend to show comparatively fewer and longer duration

565

velocity pulses (Fig. 8). The velocity pulse patterns are irregular, i.e., non-periodic. Umax does

566

not vary markedly between cases with different incidence angle configurations. For example,

567

-0.035 ~ 0.05 m s-1 in Experiment S20°IN75° and -0.04 ~ 0.03 m s-1 in Experiment S20°IN15° .

568

569

Temporal variability of flow direction at the near-bed region

570

Temporal variability of the flow velocity vector (based on streamwise and cross-stream

571

velocity, i.e., projected in the horizontal basin-floor plane) of the current recorded at 0.01 m

572

above the basin floor and/or the slope surface is examined for each ramp experiment (Figs 9-

573

12). A specific height of 0.01 m was chosen, to avoid any possible noise-induced interference,

574

whilst focusing on the near-bed velocity as this is critical for sediment transport and deposition

575

processes.

576

577This is a non-peer reviewed preprint submitted to EarthArxiv

Fig. 9. Compass plots illustrating the spatial and temporal variability of the flow velocity vector

578

(projected in the horizontal basin-floor) of the current within the quasi-steady phase (34 ~ 120

579

s) recorded at 0.01 m above the basin floor and/or the slope surface in Experiment S20°IN75°.

580

‘bc’ denotes the measurements at the base of slope along the channel-basin centreline and ‘ml’,

581

‘mc’ and ‘mr’ denote the measurements at the left, central and right flow front positions (in the

582

flow direction), respectively (ADV4, ADV3 and ADV2 in Figure 1). In each compass plot, the

583

arrow length denotes the velocity magnitude, and the direction denotes the velocity direction

584

relative to the basin. Each arrow is colour coded as time. Black dashed line indicates the slope

585

orientation. For presentation purposes, in each compass plot, the original 100 Hz ADV velocity

586

data are decimated to 10 Hz.

587

588

Flow directions at the quasi-steady phase (34 ~ 120 s)

589

Measurements during the quasi-steady phase of the current (Figs 9-11) indicate that all ramp

590

experimental configurations record complex patterns of flow direction and magnitude,

591

including the presence of multidirectional combined flow regimes above the slope surface and

592

near the base of slope.

593

For the ramp experiments (Fig. 9), flow velocity is higher at the base of slope than that at the

594

flow front positions above the slope surface (e.g., maximum velocity of ca. 0.09 m s-1 vs. ca.

595

0.05 m s-1 in Experiment S20°IN75°). Current directions recorded at the flow front positions

596

all exhibit a radial dispersal pattern whilst those recorded at the base of slope along the channel-

597

basin centreline demonstrate diverse dispersal patterns including a radial dispersal and more

598

unidirectional distribution pattern (Figs 9-11, see the descriptions below). In a single slope

599

configuration (e.g., Experiment S20°IN75°), downstream current data above the slope typically

600

601This is a non-peer reviewed preprint submitted to EarthArxiv

602

Fig. 10. Compass plots illustrating the temporal variability of the flow velocity vector

603

(projected in the horizontal basin-floor) of the current recorded at 0.01 m above the basin floor

604

and/or the slope surface within the quasi-steady phase (34 ~ 120 s) in Experiments S20°IN90°

605

(A, E), S20°IN75° (B, F), S20°IN45° (C, G) and S20°IN15° (D, H). ‘bc’ denotes the

606

measurements at the base of slope and ‘mc’ denotes the measurements at the central flow front

607

position (ADV3 in Figure 1). In each compass plot, the arrow length denotes the velocity

608

magnitude, and the direction denotes the velocity direction relative to the basin. Each arrow is

609

colour coded as time. Black dashed line indicates the slope orientation. For presentation, in

610

each compass plot, the original 100 Hz ADV velocity data are decimated to 10 Hz. See Figure

611

9 for the legend of this figure.

612

613

show an increased unidirectional component in flow direction distribution, compared to those

614

recorded upstream (reverse flow; e.g., Fig. 9A, C).

615

Across experiments with different flow incidence angles onto the slope (Fig. 10), base of slope

616

flow directions show a gradual transition from a radial to a more unidirectional dispersal pattern

617

(oriented to the along-strike direction parallel to the slope) as the flow incidence angle

618

decreases (Fig. 10E-H; 0° ~ 360° in Experiment S20°IN90° vs. 320° ~ 30° clockwise in

619This is a non-peer reviewed preprint submitted to EarthArxiv

Experiment S20°IN15°). On the slope, the unidirectional component of the flow recorded at

620

the central flow front position increases with a lower incidence angle, although all

621

configurations exhibit a radial dispersal pattern (Fig. 10A-D). However, the overall radial

622

dispersal pattern above the slope surface is established in different ways. The flow direction in

623

a highly oblique experimental configuration predominantly rotates with time, whereas in a less

624

oblique experiment the flow velocity direction tends to maintain a radial pattern through time.

625

Across experiments with different slope gradients (Fig. 11), the velocity magnitude and the

626

flow direction distribution do not vary markedly. Notably, with a steeper slope gradient, the

627

velocity magnitude recorded at the base of slope or near the flow front tends to be slightly

628

larger. Furthermore, for steeper slopes, typically the current data exhibit a slightly wider spread

629

in both overall flow directions throughout the experiment (290° ~ 15° clockwise in Experiment

630

S20°IN45° vs. 290° ~ 30° clockwise in Experiment S40°IN45°) and flow directions over a

631

given period, compared to gentler topographic slopes.

632

In summary, the incidence angle of the current relative to the containing slope appears to

633

influence the temporal variability of the flow direction at the near-bed region more strongly

634

than the slope gradient. This holds true both for the flow at the base of slope and the flow front

635

position along the channel-basin centreline.

636

637This is a non-peer reviewed preprint submitted to EarthArxiv

638

Fig. 11. Compass plots illustrating the temporal variability of the flow velocity vector

639

(projected in the horizontal basin-floor) of the current within the quasi-steady phase (34 ~ 120

640

s) recorded at 0.01 m above the basin floor and/or the slope surface in Experiments S20°IN45°

641

(A, D), S30°IN45° (B, E) and S40°IN45° (C, F). ‘bc’ denotes the measurements at the base of

642

slope and ‘mc’ denotes the measurements at the central flow front position (ADV3 in Figure

643

1). In each compass plot, the arrow length denotes the velocity magnitude, and the direction

644

denotes the velocity direction relative to the basin. Each arrow is colour coded as time. Black

645

dashed line indicates the slope orientation. For presentation, in each compass plot, the original

646

100 Hz ADV velocity data are decimated to 10 Hz. See Figure 9 for the legend of this figure.

647

648

Flow directions at the waning phase (160 ~ 180 s)

649

Temporal variability of the near-bed velocity vector above the slope surface during the waning

650

phase of the current (Fig. 12) is analysed. This stage is critical for sediment deposition process,

651This is a non-peer reviewed preprint submitted to EarthArxiv

especially the development of tractional bedforms such as ripples in the Bouma C division,

652

which in field studies are compared to sole structure orientation to interpret the presence and

653

orientation of seabed topography (e.g. Kneller et al., 1991; Hodgson and Haughton, 2004). This

654

specific time window (160 ~ 180 s), where velocities are about 10-20% of that of the quasi-

655

steady flow (Fig. 12), is chosen to avoid the later effects of reflections from the tank sidewalls.

656

Results indicate that within a near frontal experimental configuration (S20°IN75° and

657

S20°IN90°; Fig. 12G-K), the near-bed velocity vectors on the slope surface tend to be

658

dominated by a downslope flow direction with a nearly orthogonal angle to the topographic

659

slope orientation. This is likely because when the dilute flow declines higher up on the slope

660

surface, gravity starts to dominate and therefore the flow collapses orthogonal to the slope. In

661

a highly oblique or oblique experimental configuration (S20°IN15°; S20°IN45°; Fig. 12A-F),

662

the near-bed flow directions during the waning phase are more variable, with flows showing a

663

high degree of radial spreading in places (Fig. 12B, 12E and 12F), and mean flow angles in

664

the range of ~30-45 relative to the slope. This is attributed to the input flow not riding up the

665

slope as high, and therefore gravity has a minor influence relative to the basinward flow

666

momentum.

667This is a non-peer reviewed preprint submitted to EarthArxiv

668

Fig. 12. Compass plots illustrating the temporal variability of the flow velocity vector

669

(projected in the horizontal basin-floor) of the current within the waning phase (160 ~ 180 s)

670

recorded at 0.01 m above the slope surface in Experiments S20°IN15° (A-C), S20°IN45° (D-

671

F), S20°IN75° (G-I) and S20°IN90° (J, K). ‘ml’, ‘mc’ and ‘mr’ denote the measurements at the

672

left, central and right flow front positions (in the flow direction), respectively (ADV4, ADV3

673This is a non-peer reviewed preprint submitted to EarthArxiv

and ADV2 in Figure 1). In each compass plot, the arrow length denotes the velocity magnitude,

674

and the direction denotes the velocity direction relative to the basin. Each arrow is colour coded

675

as time. Black dashed line indicates the slope orientation. For presentation, in each compass

676

plot, the original 100 Hz ADV velocity data are decimated to 10 Hz. Note the different velocity

677

scale for the arrows relative to Figures 9-11.

678

679

DISCUSSION

680

Absence of internal waves in unconfined density current interactions with topographic

681

slopes

682

In all the ramp experimental configurations, no well-defined internal wave-like features are

683

observed (Videos 1-4), suggesting that features including solitons and bores do not develop

684

above all of the planar topographic slopes. This is at odds with the presence of internal waves

685

observed in previous narrow 2D flume tank (e.g., Pantin and Leeder, 1987; Edwards et al.,

686

1994; Patacci et al., 2015) and qualitative 3D experiments (Kneller et al., 1991; Haughton,

687

1994; Kneller, 1995) when density currents encounter topographic slopes. The internal waves

688

were either reflected bores or waves running along at the top of the density flow due to the

689

reflection of the currents against topographic slopes (e.g., Pantin and Leeder, 1987; Edwards

690

et al., 1994; Kneller et al., 1991) or linked to initial inlet properties of the flow such as Kelvin-

691

Helmholtz instabilities (e.g., Patacci et al., 2015). The possible explanation for the absence of

692

internal waves in this work is detailed in the following section.

693

694

Revisiting the paradigm of flow deflection and reflection

695

The prevailing paradigm for sediment gravity flow interaction with topographic slopes is that

696

flow reflection is always orthogonal to the slope irrespective of the incidence angle of the flow

697This is a non-peer reviewed preprint submitted to EarthArxiv

(Kneller et al., 1991; Kneller, 1995; Kneller and McCaffrey, 1999; Fig. 13A; note though that

698

the single experiment in Haughton (1994) is slightly anomalous). This leads to a model where

699

sole marks, representing basal conditions, can be at high angles to ripple directions, within the

700

same bed; for flows parallel with containing topography, the angle is 90 (Kneller et al., 1991;

701

Kneller, 1995; Fig. 13B). In turn, the reflections are linked to internal waves and/or solitons

702

(Pantin and Leeder, 1987; Kneller et al., 1991; Edwards et al., 1994; Haughton, 1994; Kneller,

703

1995). However, the experiments herein do not support this model with a notable absence of

704

downslope reflection at more oblique incident angles (15 and 45) during the main body of

705

the flow (Figs 10 and 11, Video 3), along with a lack of evidence for internal waves. In the

706

present experiments the dominant flow processes transition from lateral divergence-dominated,

707

through reflection-dominated, to deflection-dominated as the flow incidence angle varies from

708

90° -15° and the slope gradient changes from 20° -40° (Fig. 14).

709

710

Fig. 13. Existing process models for flow deflection and reflection when sediment gravity flows

711

encounter a topographic slope (A and B) and for the resulting relationship between sole mark

712

and ripple directions (B). In these models, flow reflections are always orthogonal to the

713

topographic slope, irrespective of the incidence angle of the flow against the slope. Ripples are

714

formed as the product of internal waves travelling on the upper interface of the gravity current,

715

as shown in (B). (C) Small-scale experiment of Kneller et al. (1991) as seen in planform,

716

showing expanding flow interacting with a slope (marked in grey). Whilst the slope is oblique

717

relative to the axial flow direction of the current, due to expansion the local flow direction is

718This is a non-peer reviewed preprint submitted to EarthArxiv

orthogonal to the slope at the point where the flow interacts with the slope.

719

720

721

Fig. 14. Schematic diagram illustrating the influence of flow incidence angle onto the

722

containing slope (A, D-F) and slope gradient (A-C) on the general flow behaviour.

723

724

The existing paradigm was developed from qualitative 3D experiments against oblique, and

725

parallel to flow, containing slopes (Kneller et al., 1991; Kneller, 1995), which therefore appear

726

paradoxical compared to the present experiments. The key to this conundrum is that the

727

previous experiments were run in a very small tank, 1 m by 1 m in planform, and consequently

728

flows were in a strongly expansional phase having exited the inlet channel when they interacted

729

with the containing slope (Kneller et al., 1991, Fig. 13C). Hence, the local flow direction

730

relative to the slope was approximately orthogonal (Kneller et al., 1991, Fig. 13C; Kneller,

731This is a non-peer reviewed preprint submitted to EarthArxiv

1995, his fig. 13). Consequently, the slopes were not oblique relative to the local flow direction

732

of the impinging flow, and therefore the resulting reflections were essentially orthogonal to the

733

slope, and thus comparable with 2D experiments on orthogonal slopes (e.g., Edwards et al.,

734

1994).

735

The previous 3D experiments (Kneller et al., 1991; Kneller, 1995) did generate clear internal

736

waves, as also observed for 2D slopes (Edwards et al., 1994), which were not observed in the

737

present experiments. Key to this difference may be the orders of magnitude differences in the

738

density of the impinging flows. In the present study, flows were dilute (~0.3% density

739

difference), in contrast to 6.7-12.8% density differences reported in Kneller et al. (1991), and

740

3% in Kneller (1995); note that these are initial values for the Kneller et al. (1991) and Kneller

741

(1995) cases, however the small tank size limited the time for entrainment and dilution prior to

742

impacting the slope. Flows that are 1-2 orders of magnitude greater in density will be prone to

743

far stronger flow reflection, and will lack the run-up heights and more complex interaction with

744

slopes observed herein. Whilst the bulk flow density of natural turbidity currents remains

745

poorly known, the best estimates range from <0.1% to ~0.2% (Konsoer et al., 2013; Simmons

746

et al., 2020), comparable to natural saline-driven density currents (~0.1-0.2%; Sumner et al.,

747

2014; Azpiroz-Zabala et al., 2024). Consequently, the present experiments are far more

748

comparable to those estimated from natural systems. However, this comparative exercise does

749

suggest that flow density is a key variable that requires further assessment.

750

The model of ripple formation from internal waves is itself problematic. This is because the

751

internal waves are postulated to form at the upper interface of the turbidity current (Kneller et

752

al., 1991; Kneller, 1995). Given that natural unconfined or partially confined turbidity currents

753

can be metres to tens of metres in thickness (e.g., Stevenson et al., 2013; Lintern et al., 2016;

754

Hill and Lintern, 2022), it is unclear if the internal waves are able to penetrate to the bed.

755

Furthermore, the internal wave driven model of Kneller (1995; Fig. 13B) has both the axial

756This is a non-peer reviewed preprint submitted to EarthArxiv

flow and the ripple generating transverse flows present at the same time. However, there is a

757

temporal gap between the formation of the sole marks and the ripples, particularly as there may

758

be a substantial time gap between the cutting of the sole marks and the deposition of the

759

immediately overlying sediment (Peakall et al., 2020; Baas et al., 2021). Furthermore, the

760

ripples in the Bouma C division are typically formed right at the end of sand deposition. Thus,

761

it could be hypothesised that the ripples may reflect the waning phase of the flow where the

762

incident flow declines, leaving gravity to dominate, with flows collapsing orthogonal to the

763

slope. For high incidence angle slopes (75 and 90) the present experiments show that waning

764

flows on slopes are orthogonal (Fig. 12G-K). In contrast, highly oblique slopes (15) and

765

oblique slopes (45) show far greater variability in flow directions in the waning flows (Fig.

766

12A-F), with flows showing a high degree of radial spreading in places (Fig. 12B), and mean

767

flow angles in the range of ~30-45 relative to the slope, rather than orthogonal (Fig. 12A-C).

768

So even waning flows in highly oblique systems are not predominantly orthogonal to slopes as

769

suggested in the existing model (Kneller et al., 1991; Kneller, 1995; Kneller and McCaffrey,

770

1999).

771

A further conundrum is that palaeocurrent data in elongate basins typically show high angles

772

between basin axial sole structures and basin transverse ripples in flows that were postulated

773

to be broadly parallel to slopes (e.g., Cope, 1959; Craig and Walton, 1962; Prentice, 1962;

774

Kelling, 1964; Seilacher and Meischner, 1965; Scott, 1967; Kneller et al., 1991; Smith and

775

Anketell, 1992), with Kneller et al. (1991) showing a peak in angular discordance between 60

776

and 90

. These field data are thus in agreement with the Kneller et al. (1991) model of

777

orthogonal reflection. Given, the experiments herein demonstrate that orthogonal reflection is

778

not universal, as previously postulated (Kneller et al., 1991), and does not occur under highly

779

oblique incidence angles, why do flow parallel field examples appear to show orthogonal flow

780

reflection? In order to address this enigma, a flow visualisation experiment was undertaken of

781This is a non-peer reviewed preprint submitted to EarthArxiv

a flow travelling parallel to a topographic ramp (Fig. 15). The visualisation (see Fig. 15 and

782

Video 5) shows that a flow that is parallel to a planar bounding surface produces a series of

783

flow fronts that move up and down the topographic ramp. Given that the incidence angle is 0

784

the flow collapses down the slope purely under gravity forcing, and thus moves orthogonal to

785

the slope. These orthogonal flows on the slope thus explain the field data from elongate basin-

786

fills.

787This is a non-peer reviewed preprint submitted to EarthArxiv

788

Fig. 15. Example images looking upstream depicting the temporal evolution of density currents

789

upon incidence with a flow-parallel topographic slope of 10° slope gradient. t denotes the

790

experimental time since the release of the flow from the mixing tank. Dye injection on the slope

791

is used to visualise the flow behaviour. Note the repeated flow-front growth and collapse above

792

the topographic slope moving in an orthogonal direction to the slope, with localised rugosity

793This is a non-peer reviewed preprint submitted to EarthArxiv

along the flow front (also see Video 5 for more detail of this flow behaviour).

794

795

796

Video 5. Annotated video illustrating the behaviour of density currents upon incidence with a

797

flow-parallel topographic slope of 10° slope gradient.

798

799

In summary, flows that are at very high angles to topographic slopes, produce orthogonal

800

reflections down the slope. As flows become more oblique, they are deflected rather than

801

reflected, and do not exhibit orthogonal reflections, even in the case of waning flows that might

802

be expected to generate ripples. Once flows become parallel to topographic slopes (incidence

803

angle of 0), however, they exhibit flow-front growth and collapse on their flank against the

804

bounding topographic slope. The collapsing flows on the flank thus are driven purely by gravity

805

and show orthogonal flow directions relative to the slope, in agreement with the palaeocurrent

806

data from elongate basin-fills. This new model of flow reflection, and deflection (Fig. 16A;

807

Fig. 14), shows that the incidence angle of the flow against the slope is critical. Flows do not

808

universally reflect orthogonally as believed for the past three decades (Kneller et al., 1991;

809

Kneller and McCaffrey, 1999). The mechanics observed herein, are also radically different to

810

that proposed in the current paradigm. Ripples are formed on slopes, and close to the base of

811

slopes, by flows moving down the slope, in many cases during the waning of flows, rather than

812This is a non-peer reviewed preprint submitted to EarthArxiv

being the product of internal waves travelling on the upper interface of the gravity current

813

(Kneller et al., 1991; Kneller, 1995; Fig. 16A-D). The present model suggests that

814

palaeocurrents showing high angles between sole marks and ripples, are formed on, or close

815

to, slopes in contrast to the model of Kneller (1995; Fig. 13B) that shows such relationships

816

occurring across entire basins.

817

818

Fig. 16. A new process model proposed in this work highlighting the importance of incidence

819

angle of the flow against the slope, on flow reflection and deflection. Flows that are at very

820

high angles to topographic slopes (A and B), produce orthogonal reflections down the slope.

821

As flows become more oblique (A and C), they are deflected rather than reflected, and do not

822

exhibit orthogonal reflections, even in the case of waning flows that might be expected to

823

generate ripples. Once flows become parallel to topographic slopes (incidence angle of 0; A

824

and D), however, they exhibit flow-front growth and collapse on their flank against the

825

bounding topographic slope. The collapsing flows on the flank thus are driven purely by gravity

826This is a non-peer reviewed preprint submitted to EarthArxiv

and show orthogonal flow directions relative to the slope. In (B-D), ripples are formed on

827

slopes, and close to the base of slopes, by flows moving down the slope, in many cases during

828

the waning of flows, rather than being the product of internal waves travelling on the upper

829

interface of the gravity current, as shown in Figure 13B.

830

831

Velocity pulsation on slopes

832

The input flow in the experiments is quasi-steady in nature (Table 1). However, distinct

833

temporal velocity pulsing, or velocity unsteadiness, in the basal part of the flows is recorded in

834

all experimental configurations, both at the base of, and on the topographic slope, as measured

835

along the channel-basin centreline (Figs 5-8). This velocity pulsing is generated by the repeated

836

fluctuations of the flow front, with periodic collapses of fluid down the slope. In turn, the nature

837

of the velocity pulsing in terms of velocity amplitude and frequency varies as a function of

838

incidence angle and slope angle; see Fig. 17 for a schematic illustration of these variations.

839

This mechanism for velocity pulsing is therefore tied to slopes and the base of slopes, but will

840

likely not propagate much farther into the basin. Slopes have previously been associated with

841

the generation of velocity pulsing, but this has either been in the form of solitons and internal

842

waves (Kneller et al., 1991, 1997; Edwards et al., 1994; Kneller, 1995; Patacci et al., 2015), or

843

the generation of true oscillatory flows has been postulated (Tinterri, 2011; Tinterri and Muzzi

844

Magalhaes, 2011). The present experiments do not show any evidence for the generation of

845

oscillatory flows, with the pulsation related to movement of fluid up and down the slope, rather

846

than propagation of a wave through the medium. Similarly, there is no evidence for solitons or

847

internal waves in the present experiments. The three-dimensional nature of the present

848

experiments and flow density values that are orders of magnitude lower than some previous

849

experiments and more commensurate with those of natural flows, likely account for the absence

850

of these solitons and internal waves, as discussed previously.

851This is a non-peer reviewed preprint submitted to EarthArxiv

852

Fig. 17. Schematic diagram illustrating the influence of different containing topographic

853

configurations (orientation and slope gradient) on the temporal pulsing pattern of the down-dip

854

velocity and temporal variability in the velocity vector (based on streamwise and cross-stream

855

velocity). As the incidence angle decreases (A and C), velocity pulsing recorded at the base of

856

slope is characterized by: i) a marked decrease in the magnitude of the maximum velocity Umax,

857

ii) a greater number of velocity pulses, and iii) a much shorter duration of each pulse. In cases

858

with a steeper slope gradient (A and B), a subtle decrease in Umax, and relatively more and

859

shorter velocity pulses are recorded. Velocity pulsing recorded at the flow front position in

860

experiments with a low flow incidence angle to the slope (A and C) is characterized by a more

861

irregular, non-periodic nature, comparatively fewer and longer velocity pulses. There is

862This is a non-peer reviewed preprint submitted to EarthArxiv

negligible difference in Umax, and relatively more and shorter velocity pulses for cases with a

863

steeper slope gradient (A and B).

864

865

This mechanism for velocity pulsing on slopes, might potentially be combined with velocity

866

pulsing mechanisms intrinsic to flows such as Kelvin-Helmholtz or Holmboe waves

867

(Kostaschuk et al., 2018), or internal waves (Marshall et al., 2021, 2023). Such pulsing

868

mechanisms are likely at a higher frequency (Kostaschuk et al., 2018), and thus subsidiary to

869

the slope induced pulsing. More complex velocity pulsation may be possible where the flows

870

themselves are driven by externally induced pulsation, such as Rayleigh-Taylor instabilities

871

generated in some plunging flows (Best et al., 2005; Dai, 2008; Kostaschuk et al., 2018), or via

872

other external drivers such as roll waves, storms, and wind- or tide-driven circulation, river

873

discharge events, cyclic slope failure (e.g., Syvitski and Hein, 1991; Ogston and Sternberg,

874

1999; Ogston et al., 2000; Li et al., 2001; Wright et al., 2002).

875

Flows that establish velocity pulses will change bed shear stresses and even alternate between

876

periods of sediment erosion and deposition. Therefore, complicated stratigraphic patterns can

877

develop despite quasi-steady inflows (cf. Best et al., 2005). Hence, more and shorter velocity

878

pulses for a single turbidity current event as documented in steeper or less oblique containing

879

slope settings (Fig. 17) may lead to complex patterns of sediment deposition, bypass and

880

transient erosion, and hence more intra-bed discontinuities, compared to their counterparts in

881

gentler or highly oblique containing slope settings, respectively. Furthermore, velocity pulsing,

882

and hence fluctuations in flow energy, may be manifested in the rock record with vertical

883

bedform variations when the velocity fluctuations occur across the thresholds of bedform

884

stability fields (Southard, 1991; cf. Ge et al., 2022). Alternations of different bed types

885

representing different flow regimes might occur due to temporal velocity pulsing. For instance,

886

in the rock record, contained turbidites on, or at the base of, slopes can be characterized by

887This is a non-peer reviewed preprint submitted to EarthArxiv

repetitive alternations of internal divisions, including switching between massive or dewatered

888

and laminated, laminated and convoluted, and parallel-laminated and ripple-laminated

889

divisions (e.g., Kneller and McCaffrey, 1999; Felletti, 2002; Muzzi Magalhaes and Tinterri,

890

2010). Higher frequency velocity pulsing at the base of slopes documented in a steep or lowly

891

oblique containing slope setting (Fig. 17) may result in more frequent alternations of internal

892

divisions. The specific type of the internal divisions might be different depending on the

893

magnitude of the near-bed velocity.

894

895

Generation and spatial variation of combined flows on slopes

896

Combined flows in deep-water settings are hypothesised to form as turbidity currents interact

897

with seafloor topography (Kneller et al., 1991; Edwards et al., 1994; Patacci et al., 2015;

898

Tinterri, 2011; Tinterri et al., 2016, 2022; Keavney et al., 2024). The experiments herein (Fig.

899

4, Figs 9-11 and Videos 1-4) support the generation of combined flow in 3D unconfined

900

density current above a topographic slope. This result is consistent with the findings in Keavney

901

et al. (2024) who address the interaction of unconfined density currents with a frontal (i.e., 90-

902

degree incidence angle) containing slope. The combined flow on the slope herein is generated

903

after the unidirectional parental flow transforms upon incidence with the slope into a

904

multidirectional parental flow on the slope surface, which then collapses downslope to

905

converge with the basal dense flow (Fig. 14 and Videos 1-4). The combined flow at the flow

906

front positions on the slope is therefore a combination of the newly generated multidirectional

907

outbound flow and the reflected flow downslope. Hence, with this study and Keavney et al.

908

(2024), a new mechanism is demonstrated for generating combined flows across a wide set of

909

topographic slope configurations, without the generation of internal waves as invoked by

910

previous studies (Kneller et al., 1991; Edwards et al., 1994; Patacci et al., 2015; Tinterri, 2011;

911

Tinterri et al., 2016, 2022). Furthermore, in contrast to the regular linear combined flows

912This is a non-peer reviewed preprint submitted to EarthArxiv

generated in confined 2D flume tank experiments (e.g., Pantin and Leeder, 1987; Edwards et

913

al., 1994; Kneller and McCaffrey, 1995; Kneller et al., 1997), the combined flows herein are

914

multidirectional, which should be much more common in nature where flows are free to spread

915

laterally on a topographic slope.

916

Crucially, this work (Figs 9-11) presents a broad range of multidirectional combined flows, the

917

unidirectional component of which varies markedly with different locations on a single

918

containing slope, as well as with different topographic slope configurations (both orientation

919

and slope gradient). Above a single planar slope, as the density current interacts with the

920

topography, the initial unidirectional parental flow is transformed into a strongly multi-

921

directional flow high-up on the slope. Therefore, more radial dispersal patterns in flow

922

direction distribution are noted for the flows documented at the flow front position compared

923

to those recorded at the base of slope (Fig. 9; Fig. 10A-D vs. Fig. 10E-H). A narrower spread

924

in flow directions along the slope (Fig. 9A-C) is likely because the reversing flow at the

925

downstream position tends to collapse downslope and converge with the basal flow running

926

parallel to the slope, likely leading to the establishment of combined flow with a unidirectional

927

component oriented parallel to the slope orientation. In a low flow incidence angle setting, the

928

increased unidirectional component of the flow recorded at the central flow front position high-

929

up on the slope (Fig. 10A-D) could be explained by an enhanced influence of flow deflection

930

running parallel to the slope on the flow directions; this is due to a decrease in topographic

931

containment from a near frontal to a highly oblique topographic slope setting (Fig. 14F).

932

This work demonstrates that multiple types of complex multidirectional combined flows can

933

be generated above planar topographic slopes by changing the orientation or slope angle of the

934

containing topographic slope. The interaction of density currents with non-planar seafloor

935

topography and unsteady flows in the field would favour the establishment of even more

936

complex patterns of combined flows above slopes. Therefore, there is no requirement for

937This is a non-peer reviewed preprint submitted to EarthArxiv

reflected bores or internal waves to generate complex combined flows as invoked in field

938

outcrop-based models above complex and/or non-planar topographic slopes (e.g., Tinterri,

939

2011; Tinterri et al., 2016, 2022).

940

941

A new model for deposits on orthogonal and oblique slopes

942

Formation and spatial distribution of combined flow bedforms on slopes

943

Combined flow sedimentary structures, including small- to medium-scale biconvex

944

(mega)ripples with internal sigmoidal-cross laminae, and hummock-like bedforms, have been

945

identified in deep-water turbidites at outcrop (e.g., Marjanac, 1990; Haughton, 1994; Remacha

946

et al., 2005; Mulder et al., 2009; Tinterri, 2011; Tinterri et al., 2016, 2022; Hofstra et at., 2018;

947

Martínez-Doñate et al., 2021; Privat et al., 2021; Taylor et al., 2024). The formation of these

948

sedimentary structures is typically hypothesised to be linked to generation of combined flows

949

by the superposition of a unidirectional parental turbidity current with an oscillatory component

950

due to the reflections of the internal waves or bores against a topographic slope (Tinterri, 2011;

951

Tinterri et al., 2016, 2022; see also Kneller et al., 1991; Edwards et al., 1994; Haughton, 1994),

952

largely on the basis of observations of internal waves in 2D or qualitative 3D reflected density

953

current experiments (e.g., Kneller et al., 1991; Edwards et al., 1994). Nevertheless, the present

954

experimental work documents the generation of complex, multidirectional combined flows on

955

the slope surface when unconfined turbidity currents interact with all oblique topographic slope

956

configurations (Figs 9-11; Videos 1-4). This is at odds with these previous models, and instead

957

supports the model for the formation of hummock-like bedforms through combined flows on

958

slopes as proposed by Keavney et al. (2024). Herein, this model of Keavney et al. (2024) is

959

demonstrated to be applicable in a wider range of topographic configurations, and a new

960

mechanism for sigmoidal bedforms is proposed, without requirement for an oscillatory

961

component. Hummock-like bedforms form during relatively high sediment fallout rates when

962This is a non-peer reviewed preprint submitted to EarthArxiv

flows decelerate upon incidence with the slope, and under combined flow conditions with a

963

radial dispersal pattern (Keavney et al., 2024). Sigmoidal bedforms form during relatively

964

lower sediment fallout rates, under combined flows with a radial dispersal pattern but a strong

965

unidirectional component.

966

Depending on the relative strength of the unidirectional component of the multidirectional

967

combined flow documented on slopes in this work (Figs 9-11 and Fig. 14), hummock-like

968

bedforms in these settings are expected to be characterized by various degrees of anisotropy,

969

and transition into symmetric or asymmetric biconvex ripples with internal sigmoidal laminae

970

when the unidirectional component of the combined flow increases. In a single topographic

971

slope, once the particulate density currents encounter the topography, flow decelerates, leading

972

to an increase in suspension fallout rate; the unidirectional parental flow is transformed into a

973

strongly multi-directional flow high-up on the slope. Therefore, more isotropic hummock-like

974

bedforms are predicted to form high-up on the slope under such combined flows (see also

975

Keavney et al., 2024; Fig. 18A). Along the in-flow direction high-up on a single slope, the

976

transformed multi-directional flow tends to finally collapse downslope to converge to the basal

977

flow to run parallel to the slope, and hence the combined flow along an in-flow direction tends

978

to show a progressive unidirectional component oriented parallel to the slope (Fig. 10A-C).

979

Therefore, more anisotropic hummock-like bedforms, or even sigmoidal bedforms along the

980

slope, are expected to form (Fig. 18). Lower on the slope, the superposition of the strong

981

unidirectional parental flow and reflected flow downslope may lead to the deposition of more

982

anisotropic hummock-like bedforms oriented perpendicular to or parallel to the slope

983

depending on the flow incidence angle (Fig. 18).

984

As the flow incidence angle decreases (Fig. 18A-C), the enhanced dominance of flow

985

deflection versus reflection (Fig. 14) is documented to result in a progressive increase in the

986

unidirectional component of the generated combined flows high-up on the slope (Fig. 10A-D).

987This is a non-peer reviewed preprint submitted to EarthArxiv

This in turn may lead to the deposition of hummock-like bedforms characterized by an

988

increased degree of anisotropy (isotropic to strongly anisotropic) or even sigmoidal bedforms

989

when the unidirectional component is very strong. In settings across different slope gradients

990

of the topographic slope, the hummock-like bedforms on the slope surface would not show a

991

marked difference in the degree of anisotropy due to the subtle difference in the types of the

992

generated combined flow (Fig. 11A-C). This means that the degree of anisotropy in hummock-

993

like bedforms is a good indicator of the orientation of the topographic slope, or the flow

994

incidence angle to the topographic slope, but not of the slope gradient.

995

996

Fig. 18. Schematic diagrams illustrating the model of deposits for the interaction of the 3D

997

unconfined turbidity current with different combinations of containing topographic

998

configurations, including slope gradient and orientation: (A) high-angle intrabasinal slope

999

oriented orthogonal to the incoming flow; (B) low-angle intrabasinal slope oriented nearly

1000

orthogonal to the incoming flow; (C) low-angle intrabasinal slope oriented highly oblique to

1001

the incoming flow. For each slope configuration, the predicted palaeocurrent distribution

1002This is a non-peer reviewed preprint submitted to EarthArxiv

patterns, key types of bedforms, sediment dispersal patterns and onlap styles on slopes are

1003

indicated.

1004

1005

General depositional model

1006

The flow process model described herein (Fig. 14) is most applicable to basins where the flow

1007

volume is smaller than the basin capacity (i.e., unconfined flow) and the flow interacts with

1008

high-relief intrabasinal topography with a quasi-steady input flow source. For example, syn-

1009

and early post-rift (e.g., Ravnås and Steel, 1997; Cullen et al., 2020) or oblique-slip (Hodgson

1010

and Haughton, 2004; Baudouy et al., 2021) settings where fault scarps have a pronounced

1011

seabed expression.

1012

For scenarios with a low-gradient intrabasinal slope oriented nearly perpendicular to the

1013

incoming flow (Fig. 18B), processes are dominated by divergence and reflection (Fig. 4, 14A

1014

and 14D). The initial flow is observed to decouple into two parts upon incidence of the

1015

topographic slope: basal dense region and upper dilute region. The denser basal region of the

1016

flow decelerates rapidly at the base of slope due to limited upslope momentum and would

1017

therefore lead to the deposition of coarser-grained sediment fraction lower on the slope and

1018

abrupt terminations or pinch-outs (Keavney et al., 2024). At the same time, the upper dilute

1019

part of the flow can travel higher up on the slope and thin and decelerate on the slope surface,

1020

which would result in the deposition of finer-grained sediment fraction draping higher up on

1021

the slope surface (Keavney et al., 2024). The combined flows generated above the slope surface

1022

would enhance the development of more isotropic hummock-like bedforms.

1023

For scenarios with a low-gradient intrabasinal slope oriented highly oblique to the incoming

1024

flow (Fig. 18C), the flow process is deflection-dominated with limited upslope momentum and

1025

flow-topography interaction (Video 3 and Fig. 14F). Weak flow decoupling and flow stripping

1026This is a non-peer reviewed preprint submitted to EarthArxiv

on slopes is hypothesized to result in the deposition of a limited zone of draped fines, which

1027

abruptly terminates lower on the slope. The combined flows generated above the slope surface

1028

would favour the development of more anisotropic hummock-like bedforms or even biconvex

1029

ripples with internal sigmoidal laminae oriented parallel to the slope orientation.

1030

For scenarios with an intrabasinal slope of a steeper gradient (Fig. 18A), flow is more

1031

deflection dominated (Video 1 and Fig. 14C). The decreased flow stripping on the slope

1032

surface would lead to less pronounced draping of the finer-grained sediment fraction on the

1033

slope surface compared to its gentler gradient counterpart (Fig. 18B). The rapid flow

1034

deceleration at the base of the slope would lead to high rates of suspension fall out and

1035

formation of thick coarser-grained sediment fraction, abruptly terminating lower on the slope.

1036

In this scenario, an increased relative strength between flow deflection and reflection might

1037

lead to a thinner division in sedimentary facies with evidence for flow reflections (Fig. 18A)

1038

compared to lower-gradient slopes.

1039

The depositional model herein presents the first and most detailed model so far to address the

1040

interaction of unconfined turbidity currents and containing topographic slopes. Distinct onlap

1041

styles and sedimentary facies in these topographic configurations can be used to reconstruct

1042

the orientation and slope gradient of the intrabasinal or basin bounding slopes in the ancient

1043

rock record.

1044

1045

CONCLUSIONS

1046

Large-scale 3D physical experiments are utilised to examine the interaction of unconfined

1047

density currents with planar slopes at a range of orientations and gradients, and subsequently

1048

used to present the implications of the results for sedimentation on submarine slopes. The

1049

experiments show that the dominant flow process transitions from divergence-dominated,

1050This is a non-peer reviewed preprint submitted to EarthArxiv

through reflection-dominated to deflection-dominated as the flow incidence angle varies from

1051

90° to 15° and the slope gradient changes from 20° to 40° . Patterns of near-bed velocity pulsing

1052

at the base of, and on, the slope vary as a function of both the flow incidence angle and slope

1053

gradient. In all configurations, complex multidirectional combined flows are observed on, or

1054

at the base of, the slope, the types of which are shown to vary spatially across the slope and

1055

different configurations of slopes.

1056

The findings challenge the paradigm of flow deflection and reflection in existing flow-

1057

topography process models that has stood for three decades. A new process model for flow-

1058

slope interactions is presented, which provides new mechanics for the observation of high-

1059

angular differences between sole marks and ripple directions documented in many field

1060

datasets. A new mechanism for the velocity pulsation on slopes is proposed and the

1061

documentation of different patterns of velocity pulsing on slopes across different topographic

1062

configurations is presented to attribute to the formation of distinctive stratigraphic patterns in

1063

the rock record. The generation and spatial distribution of multiple types of complex

1064

multidirectional combined flows on oblique slopes further supports the generation of combined

1065

flow in 3D unconfined density current above a topographic slope, in the absence of internal

1066

waves or solitons. Specifically, the unidirectional component of the combined flows varies

1067

spatially on a slope, as well as with different topographic configurations. This process model

1068

provides a novel mechanism for the formation of different types of combined-flow bedforms

1069

on a slope and across different slope configurations in deep-sea settings.

1070

The new models of the generation and spatial distribution of combined flows and velocity

1071

pulsation patterns, coupled with sediment dispersal patterns and onlap styles on slopes provide

1072

an improved model of turbidity current sedimentation on slopes, which can be applied to refine

1073

interpretations of exhumed successions. Nonetheless, given the complicated process responses

1074

arising from simple topographic configurations documented herein, there remains much to

1075This is a non-peer reviewed preprint submitted to EarthArxiv

learn about the interactions of sediment gravity flows and seabed relief, and their depositional

1076

expression.

1077

1078

ACKNOWLEDGEMENTS

1079

This research forms a part of the LOBE 3 consortium project, based at University of Leeds and

1080

University of Manchester. The authors thank the sponsors of the LOBE 3 consortium project

1081

for financial support: Aker BP, BHP, BP, Equinor, HESS, Neptune, Petrobras, PetroChina, Total,

1082

Vår Energi and Woodside.

1083

1084

NOMENCLATURE

1085

Hmax: Maximum run-up height (m)

1086

h: Flow height (m)

1087

Fr: Froude number

1088

Frd: Densimetric Froude number

1089

g: Acceleration due to gravity (m s-2)

1090

g': Reduced gravitational acceleration (m s-2)

1091

hp: Height of the maximum downstream velocity above the basin floor (m)

1092

Re: Reynolds number

1093

t: Experimental time since the release of the flow from the mixing tank (s)

1094

U: Mean depth-averaged downstream velocity (m s-1)

1095

Umax: Maximum velocity over height on the time series profiles of down-dip velocity (m s-1)

1096

u: Streamwise velocity or down-dip velocity (m s-1)

1097

up: Maximum downstream velocity (m s-1)

1098

v: Cross-stream velocity or along-strike velocity (m s-1)

1099

w: Vertical velocity (m s-1)

1100This is a non-peer reviewed preprint submitted to EarthArxiv

 : Dynamic viscosity (Pa s)

1101

ρa : Density of the ambient fluid (kg m-3)

1102

ρs: Mean depth-averaged density of the current (kg m-3)

1103

1104

DATA AVAILABILITY STATEMENT

1105

The data that support the findings of this study are available from the corresponding author

1106

upon reasonable request. The high-resolution original experimental video files are publicly

1107

available and can be downloaded from the GitHub Repository: https://leeds365-

1108

my.sharepoint.com/:p:/g/personal/earrwa_leeds_ac_uk/EXyljFoj0GZBuIQHux7-

1109

dVEBvbqChhhejDVD-F-_QG0Ppw?e=KjyfSJ.

1110

1111

REFERENCES

1112

Alexander, J. and Morris, S. (1994) Observations on experimental, nonchannelized, high-

1113

concentration turbidity currents and variations in deposits around obstacles. J. Sediment. Res.,

1114

64, 899-909.

1115

Amy, L.A., McCaffrey, W.D. and Kneller, B.C. (2004) The influence of a lateral basin-slope

1116

on the depositional patterns of natural and experimental turbidity currents. In: Deep-water

1117

sedimentation in the Alpine foreland basin of SE France: New perspectives on the Grès

1118

d’Annot and related systems (Eds P. Joseph and S.A. Lomas), Geol. Soc. London. Spec. Publ.,

1119

221, 311-330.

1120

Athmer, W., Groenenberg, R.M., Luthi, S.M., Donselaar, M.E., Sokoutis, D. and

1121

Willingshofer, E. (2010) Relay ramps as pathways for turbidity currents: a study combining

1122

analogue sandbox experiments and numerical flow simulations. Sedimentology, 57, 806-823.

1123This is a non-peer reviewed preprint submitted to EarthArxiv

Azpiroz-Zabala, M., Sumner, E.J., Cartigny, M.J.B., Peakall, J., Clare, M.A., Darby,

1124

S.E., Parsons, D.R., Dorrell, R.M., Özsoy, E., Tezcan, D., Wynn, R.B. and Johnson, J.

1125

(2024) Benthic biology influences sedimentation in submarine channel bends: Coupling of

1126

biology, sedimentation and flow. The Depositional Record, 10, 159-175, doi:

1127

10.1002/dep2.265.

1128

Baas, J.H., Tracey, N.D. and Peakall, J. (2021) Sole marks reveal deep-marine depositional

1129

process and environment: Implications for flow transformation and hybrid event bed models.

1130

Journal of Sedimentary Research, 91, 986–1009.

1131

Bakke, K., Kane, I.A., Martinsen, O.J., Petersen, S.A., Johansen, T.A., Hustoft, S.,

1132

Jacobsen, F.H. and Groth, A. (2013) Seismic modeling in the analysis of deep-water

1133

sandstone termination styles. AAPG Bull., 97, 1395-1419.

1134

Baudouy, L., Haughton, P.D. and Walsh, J.J. (2021) Evolution of a fault-controlled, deep-

1135

water Sub-Basin, Tabernas, SE Spain. Front. Earth Sci., 9, 767286.

1136

Bell, D., Stevenson, C.J., Kane, I.A., Hodgson, D.M. and Poyatos-Moré, M. (2018)

1137

Topographic controls on the development of contemporaneous but contrasting basin-floor

1138

depositional architectures. J. Sed. Res., 88, 1166-1189.

1139

Best, J.L., Kirkbride, A.D. and Peakall, J. (2001) Mean flow and turbulence structure of

1140

sediment‐laden gravity currents: new insights using ultrasonic Doppler velocity profiling. In:

1141

Particulate Gravity Currents (Eds W.D. McCaffrey, B.C. Kneller and J. Peakall), IAS Spec.

1142

Publ., 31, 159–172.

1143

Best, J.L., Kostaschuk, R.A., Peakall, J., Villard, P.V. and Franklin, M. (2005) Whole flow

1144

field dynamics and velocity pulsing within natural sediment-laden underflows. Geology, 33,

1145

765-768.

1146This is a non-peer reviewed preprint submitted to EarthArxiv

Bruschi, R., Bughi, S., Spinazzè, M., Torselletti, E. and Vitali, L. (2006) Impact of debris

1147

flows and turbidity currents on seafloor structures. Norwegian J. Geol., 86, 317–336.

1148

Buckee, C., Kneller, B. and Peakall. J. (2001) Turbulence structure in steady, solute-driven

1149

gravity currents. In: Particulate gravity currents (Eds W. McCaffrey, B. Kneller, J. Peakall),

1150

Int. Assoc. Sedimentol. Spec. Pub., 31, 173-187.

1151

Carter, L., Gavey, R., Talling, P. and Liu, J. (2014) Insights into submarine geohazards from

1152

breaks in subsea telecommunication cables. Oceanography, 27, 58–67.

1153

Cartigny, M.J., Ventra, D., Postma, G. and Van Den Berg. J.H. (2014) Morphodynamics

1154

and sedimentary structures of bedforms under supercritical-flow condition: new insights from

1155

flume experiments. Sedimentology, 61, 712-748.

1156

Chadwick, R., Zweigel, P., Gregersen, U., Kirby, G.A., Holloway, S. and Johannessen, P.

1157

(2004) Geological reservoir characterization of a CO2 storage site: The Utsira Sand, Sleipner,

1158

Northern North Sea. Energy, 29, 1371-1381.

1159

Cope, R.N. (1959) The Silurian rocks of the Devilsbit Mountain district, County Tipperary.

1160

Proc. Roy. Irish Acad., 60, 217-242.

1161

Craig, G.Y. and Walton, E.K. (1962) Sedimentary structures and palaeocurrent directions

1162

from the Silurian rocks of Kirkcudbrightshire. Trans. Edinb. Geol. Soc., 19, 100–119.

1163

Cullen, T.M., Collier, R.E.L., Gawthorpe, R.L., Hodgson, D.M. and Barrett, B.J. (2020)

1164

Axial and transverse deep‐water sediment supply to syn‐rift fault terraces: Insights from the

1165

West Xylokastro Fault Block, Gulf of Corinth, Greece. Basin Res., 32, 1105-1139.

1166

Dai, A. (2008) Analysis and modeling of plunging flows. PhD thesis, University of Illinois at

1167

Urbana-Champaign, Illinois, US.

1168This is a non-peer reviewed preprint submitted to EarthArxiv

Dzulynski, S., Ksiazkiewicz, M. and Kuenen, P. H. (1959) Turbidites in flysch of the Polish

1169

Carpathian Mountains. Geol. Soc. Am. Bull., 70, 1089-1118.

1170

Edwards, D.A., Leeder, M.R., Best, J.L. and Pantin, H.M. (1994) On experimental reflected

1171

density currents and the interpretation of certain turbidites. Sedimentology, 41, 437-461.

1172

Felletti, F. (2002) Complex bedding geometries and facies associations of the turbiditic fill of

1173

a confined basin in a transpressive setting (Castagnola Fm., Tertiary Piedmont Basin, NW

1174

Italy). Sedimentology, 49, 645-667.

1175

Ge, Z., Nemec, W., Vellinga, A.J. and Gawthorpe, R.L. (2022) How is a turbidite actually

1176

deposited? Sci. Adv., 8, eabl9124.

1177

Gilbert, R. (1975) Sedimentation in Lillooet Lake. Can. J. Earth Sci., 12, 1697–1711.

1178

Haughton, P.D. (1994) Deposits of deflected and ponded turbidity currents, Sorbas Basin,

1179

Southeast Spain. J. Sediment. Res., 64, 233-246.

1180

Haughton, P.D. (2000) Evolving turbidite systems on a deforming basin floor, Tabernas, SE

1181

Spain. Sedimentology, 47, 497-518.

1182

Haward, M. (2018) Plastic pollution of the world’s seas and oceans as a contemporary

1183

challenge in ocean governance. Nat. Commun., 9, 667.

1184

Hill, P.R. and Lintern, D.G. (2022) Turbidity currents on the open slope of the Fraser Delta.

1185

Marine Geology, 445, 106738.

1186

Ho, V.L., Dorrell, R.M., Keevil, G.M., Thomas, R.E., Burns, A.D., Baas, J.H. and

1187

McCaffrey, W.D. (2019) Dynamics and deposition of sediment-bearing multi-pulsed flows

1188

and geological implication. J. Sediment. Res., 89, 1127-1139.

1189This is a non-peer reviewed preprint submitted to EarthArxiv

Hodgson, D.M. and Haughton, P.D. (2004) Impact of syndepositional faulting on gravity

1190

current behaviour and deep-water stratigraphy: Tabernas-Sorbas Basin, SE Spain. In: Confined

1191

Turbidite Systems (Eds S.A. Lomas, P. Joseph), Geol. Soc. Spec. Publ., 222, 135-158.

1192

Hofstra, M., Peakall, J., Hodgson, D.M. and Stevenson, C.J. (2018) Architecture and

1193

morphodynamics of subcritical sediment waves in ancient channel-lobe transition zone.

1194

Sedimentology. 65, 2339-2367.

1195

Howlett, D.M., Ge, Z., Nemec, W., Gawthorpe, R.L., Rotevatn, A. and Jackson, C.A-L.

1196

(2019) Response of unconfined turbidity currents to deep-water fold and thrust belt

1197

topography: Orthogonal incidence on solitary and segmented folds. Sedimentology, 66, 2425-

1198

2454.

1199

Kane, I.A., Clare, M.A., Miramontes, E., Wogelius, R., Rothwell, J.J., Garreau, P. and

1200

Pohl, F. (2020) Seafloor microplastic hotspots controlled by deep-sea circulation. Science, 368,

1201

1140-1145.

1202

Keavney, E., Peakall, J., Wang, R., Hodgson, D.M., Kane, I.A., Keevil, G.M., Brown,

1203

H.C., Clare, M.A. and Hughes M. (2024) Flow evolution and velocity structure of unconfined

1204

density currents interacting with frontally containing slopes. EarthArxiv, doi:

1205

10.31223/X5CM35.

1206

Keevil, G.M., Peakall, J., Best, J.L. and Amos, K.J. (2006) Flow structure in sinuous

1207

submarine channels: Velocity and turbulence structure of an experimental submarine channel.

1208

Mar. Geol., 229, 241-257.

1209

Kelling, G. (1964) The turbidite concept in Britain. In: Turbidites (Developments in

1210

sedimentology, Volume 3) (Eds A.H. Bouma and A. Brouwer), pp. 75-92. Elsevier, Amsterdam.

1211This is a non-peer reviewed preprint submitted to EarthArxiv

Kneller, B. (1995) Beyond the turbidite paradigm: physical models for deposition of turbidites

1212

and their implications for reservoir prediction. In: Characterisation of Deep Marine Clastic

1213

Systems (Eds A.J. Hartley and D.J. Prosser), Geol. Soc. London. Spec. Publ., 94, 31-49.

1214

Kneller, B. and Buckee, C. (2000) The structure and fluid mechanics of turbidity currents: a

1215

review of some recent studies and their geological implications. Sedimentology, 47, 62-94.

1216

Kneller, B.C. and McCaffrey, W.D. (1995) Modelling the effects of salt-induced topography

1217

on deposition from turbidity currents. In: Salt, sediment and hydrocarbons (Eds C.J. Travis,

1218

B.C Vendeville, H. Harrison, F.J. Peel, M.R. Hudec, B.E. Perkins), SEPM Soc. Sediment Geol.,

1219

137-145.

1220

Kneller, B. and McCaffrey, W. (1999) Depositional effects of flow nonuniformity and

1221

stratification within turbidity currents approaching a bounding slope: deflection, reflection, and

1222

facies variation. J. Sed. Res., 69, 980-991.

1223

Kneller, B., Edwards, D., McCaffrey, W. and Moore, R. (1991) Oblique reflection of

1224

turbidity currents. Geology, 14, 250-252.

1225

Kneller, B.C., Bennett, S.J., and McCaffrey, W.D. (1997) Velocity and turbulence structure

1226

of density currents and internal solitary waves: potential sediment transport and the formation

1227

of wave ripples in deep water. Sed. Geol., 112, 235-250.

1228

Konsoer, K., Zinger, J. and Parker, G. (2013) Bankfull hydraulic geometry of submarine

1229

channels created by turbidity currents: Relations between bankfull channel characteristics and

1230

formative flow discharge. Journal of Geophysical Research: Earth Surface, 118, 216–228,

1231

doi:10.1029/2012JF002422.

1232This is a non-peer reviewed preprint submitted to EarthArxiv

Kostaschuk, R., Nasr‐Azadani, M.M., Meiburg, E., Wei, T., Chen, Z., Negretti, M.E.,

1233

Best, J., Peakall, J. and Parsons, D.R. (2018) On the causes of pulsing in continuous turbidity

1234

currents. J. Geophys. Res. Earth Surf., 123, 2827-2843.

1235

Kuenen, P.H. and Migliorini, C.I. (1950) Turbidity currents as a cause of graded bedding. J.

1236

Geol., 58, 91–127.

1237

Li, G., Tang, Z., Yue, S., Zhuang, K., and Wei, H. (2001) Sedimentation in the shear front

1238

off the Yellow River mouth. Cont. Shelf Res., 21, 607–625.

1239

Lintern, D.G., Hill, P.R. and Stacey, C. (2016) Powerful unconfined turbidity current

1240

captured by cabled observatory on the Fraser River delta slope, British Columbia, Canada.

1241

Sedimentology, 63, 1041-1064.

1242

Lloyd, C., Huuse, M., Barrett, B.J. and Newton, A.M.W. (2021) Regional exploration and

1243

characterisation of CO2 storage prospects in the Utsira-Skade aquifer, North Viking graben,

1244

North Sea. Earth Sci. Syst. Soc., 1, 10041.

1245

Lomas, S.A. and Joseph, P. (2004) Confined turbidite systems. In: Confined Turbidite Systems

1246

(Eds S.A. Lomas, P. Joseph), Geol. Soc. Spec. Publ., 222, 1-7.

1247

Lusseyran, F., Izrar, B., Audemar, C. and Skali-lami, S. (2003) Time-space characteristics

1248

of stratified shear layer from UVP measurements. Exp. Fluids, 35, 32–40.

1249

Marjanac, T. (1990) Reflected sediment gravity flows and their deposits in flysch of Middle

1250

Dalmatia, Yugoslavia. Sedimentology, 37, 921-929.

1251

Marshall, C.R., Dorrell, R.M., Keevil, G.M., Peakall, J. and Tobias, S.M. (2021)

1252

Observations of large-scale coherent structures in gravity currents: implications for flow

1253

dynamics. Exp. Fluids, 62, 120.

1254This is a non-peer reviewed preprint submitted to EarthArxiv

Marshall, C.R., Dorrell, R.M., Keevil, G.M., Peakall, J. and Tobias, S.M. (2023) On the

1255

role of transverse motion in pseudo-steady gravity currents. Exp. Fluids, 64, 63,

1256

doi:10.1007/s00348-023-03599-7.

1257

Martínez-Doñate, A., Privat, A.M-L., Hodgson, D.M., Jackson, C.A-L., Kane, I.A.,

1258

Spychala, Y.T., Duller, R.A., Stevenson, C., Keavney, E., Schwarz, E. and Flint, S.S.

1259

(2021) Substrate entrainment, depositional relief, and sediment capture: impact of a submarine

1260

landslide on flow process and sediment supply. Front. Earth. Sci., 9, 757617.

1261

McCaffrey, W.D. and Kneller, B.C. (2001) Process controls on the development of

1262

stratigraphic trap potential on the margins of confined turbidite systems and aids to reservoir

1263

evaluation. AAPG Bull., 85, 971-988.

1264

Middleton, G.V. (1993) Sediment deposition from turbidity currents. Annu. Rev. Earth Planet.

1265

Sci., 21, 89–114.

1266

Muck, M.T. and Underwood, M.B. (1990) Upslope flow of turbidity currents: a comparison

1267

among field observations, theory, and laboratory methods. Geology, 18, 54–57.

1268

Mulder, T., Razin, P. and Faugeres, J-C. (2009) Hummocky cross-stratification-like

1269

structures in deep-sea turbidites: Upper Cretaceous Basque basins (Western Pyrenees, France).

1270

Sedimentology, 56, 997–1015.

1271

Muzzi Magalhaes, P. and Tinterri, R. (2010) Stratigraphy and depositional setting of slurry

1272

and contained (reflected) beds in the Marnoso-Arenacea Formation (Langhian-Serravallian)

1273

Northern Apennines, Italy. Sedimentology, 57, 1685–1720.

1274

Normark, W.R., Posamentier, H. and Mutti, E. (1993) Turbidite systems: state of the art and

1275

future directions. Rev. Geophys., 31, 91-116.

1276This is a non-peer reviewed preprint submitted to EarthArxiv

Ogston, A.S. and Sternberg, R.W. (1999) Sediment-transport events on the northern

1277

California continental shelf. Mar. Geol., 154, 69–82.

1278

Ogston, A.S., Cacchione, D.A., Sternberg, A.S. and Kineke, G.C. (2000) Observations of

1279

storm and river flood-driven sediment transport on the northern California continental shelf.

1280

Cont. Shelf Res., 20, 2141–2162.

1281

Pantin, H.M. and Leeder, M.R. (1987) Reverse flow in turbidity currents: the role of internal

1282

solitons. Sedimentology, 34, 1143-1155.

1283

Patacci, M., Haughton, P.D.W. and McCaffrey, W.D. (2015) Flow behaviour of ponded

1284

turbidity currents. J. Sed. Res., 85, 885-902.

1285

Peakall, J., Ashworth, P. and Best, J. (1996) Physical modelling in fluvial geomorphology:

1286

principles, applications and unresolved issues, in The Scientific Nature of Geomorphology:

1287

proceedings of the 27th Binghamton symposium, September 27-29, 1996. (Eds B. Rhoads and

1288

C. Thorn) (Hoboken, NJ: Wiley and Sons Ltd), 221-253.

1289

Peakall, J., Best, J.L., Baas, J., Hodgson, D.M., Clare, M.A., Talling, P.J., Dorrell, R.M.

1290

and Lee, D.R. (2020) An integrated process-based model of flutes and tool marks in deep-

1291

water environments: implications for palaeohydraulics, the Bouma sequence, and hybrid event

1292

beds. Sedimentology, 67, 1601–1666, 10.1111/SED.12727.

1293

Pickering, K.T. and Hiscott, R.H. (1985) Contained (reflected) turbidity currents from the

1294

Middle Ordovician Cloridorme Formation, Quebec, Canada: an alternative to the antidune

1295

hypothesis. Sedimentology, 32, 373-394.

1296

Prentice, J.E. (1962) The sedimentary history of the Carboniferous in Devon. In: Some aspects

1297

of the Variscan fold belt (Ed. K. Coe), pp. 93-108. Manchester University Press, Manchester,

1298

England.

1299This is a non-peer reviewed preprint submitted to EarthArxiv

Privat, A.M-L.J., Hodgson, D.M., Jackson, C.A-L., Schwarz, E. and Peakall, J. (2021)

1300

Evolution from syn-rift carbonates to early post-rift deep-marine intraslope lobes: The role of

1301

rift basin physiography on sedimentation patterns. Sedimentology, 68, 2563-2605.

1302

Ravnås, R. and Steel, R.J. (1997) Contrasting styles of Late Jurassic syn-rift turbidite

1303

sedimentation: a comparative study of the Magnus and Oseberg areas, northern North Sea.

1304

Mar. Petrol. Geol., 14, 417-449.

1305

Remacha, E., Fernandez, L.P. and Maestro, E. (2005) The transition between sheet-like lobe

1306

and basin-plain turbidites in the Hecho Basin (South-Central Pyrenees, Spain). J. Sed. Res., 75,

1307

798-819.

1308

Scott, K.M. (1967) Intra-bed palaeocurrent variations in a Silurian flysch sequence,

1309

Kircudbrightshire, Southern Uplands of Scotland. Scott. J. Geol., 3, 268-281.

1310

Seilacher, A. and Meischner, D. (1965) Fazies-analyse im palaozoikum des Oslo-Gebeites.

1311

Geologische Rundschau, 54, 596-619.

1312

Sequeiros, O.E., Spinewine, B., Beaubouef, R.T., Sun, T., Garcia, M.H. and Parker, G.

1313

(2010) Bedload transport and bed resistance associated with density and turbidity

1314

currents. Sedimentology, 57, 1463-1490.

1315

Sestini G. (1970) Flysch facies and turbidite sedimentology. Sediment. Geol., 4, 559–597.

1316

Simmons, S.M., Azpiroz‐Zabala, M., Cartigny, M.J.B., Clare, M.A., Cooper, C., Parsons,

1317

D.R., Pope, E.L., Sumner, E.J. and Talling, P.J. (2020). Novel acoustic method provides first

1318

detailed measurements of sediment concentration structure within submarine turbidity currents.

1319

Journal of Geophysical Research: Oceans, 125, e2019JC015904, doi: 10.1029/2019JC015904.

1320

Sinclair, H.D. (1994) The influence of lateral basinal slopes on turbidite sedimentation in the

1321

Annot Sandstones of SE France. J. Sediment. Res., 64, 42–54.

1322This is a non-peer reviewed preprint submitted to EarthArxiv

Smith, R.D.A. and Anketell, J.M. (1992) Welsh Basin ‘contourites’ reinterpreted as fine-

1323

grained turbidites: the Grogal Sandstones. Geological Magazine, 129, 609-614.

1324

Southard, J.B. (1991) Experimental determination of bed-form stability. Annu. Rev. Earth

1325

Planet. Sci., 19, 423-455.

1326

Southern, S.J., Patacci, M., Felletti, F. and McCaffrey, W.D. (2015) Influence of flow

1327

containment and substrate entrainment upon sandy hybrid event beds containing a co-genetic

1328

mud-clast-rich division. Sediment. Geol., 321, 105-122.

1329

Soutter, E.L., Bell, D., Cumberpatch, Z.A., Ferguson, R.A., Spychala, Y.T., Kane, I.A.

1330

and Eggenhuisen, J.T. (2021) The influence of confining topography orientation on

1331

experimental turbidity currents and geological implications. Front. Earth. Sci., 8. 540633.

1332

Stevenson, C.J., Talling, P.J., Wynn, R.B., Masson, D.G., Hunt, J.E., Frenz, M.,

1333

Akhmetzhanov, A. and Cronin, B.T. (2013) The flows that left no trace: Very large-volume

1334

turbidity currents that bypassed sediment through submarine channels without eroding the sea

1335

floor. Marine and Petroleum Geology, 41, 186-205.

1336

Sumner, E.J., Peakall, J., Dorrell, R.M., Parsons, D.R., Darby, S.E., Wynn, R.B.,

1337

McPhail, S.D., Perrett, J., Webb, A., and White D. (2014) Driven around the bend: spatial

1338

evolution and controls on the orientation of helical bend flow in a natural submarine gravity

1339

current. Journal of Geophysical Research – Oceans, 119, 898-913, doi:

1340

10.1002/2013JC009008.

1341

Syvitski, J.P.M. and Hein, F.J. (1991) Sedimentology of an Artic basin: Itirbilung Fiord,

1342

Baffin Island, Northwest Territories. Geological Survey of Canada Paper, 90, 66.

1343

Takeda, Y. (1991) Development of an Ultrasound Velocity Profile Monitor. Nucl. Eng. Des.,

1344

126, 277–284.

1345This is a non-peer reviewed preprint submitted to EarthArxiv

Takeda, Y. (1993) Velocity Profile Measurement by Ultrasonic Doppler Method. In:

1346

Experimental Heat Transfer, Fluid Mechanics and Thermodynamics (Ed. M.D. Kelleher), pp.

1347

126–131. Elsevier, Amsterdam.

1348

Taylor, W.J., Hodgson, D.M., Peakall, J., Kane, I.A., Morris, E.A. and Flint, S.S. (2024)

1349

Unidirectional and combined transitional flow bedforms: Controls on process and distribution

1350

in submarine slope settings. Sedimentology, doi: 10.1111/sed.13177.

1351

Thomas, R.E., Schindfessel, L., McLelland, S.J., Creëlle, S. and De Mulder, T. (2017) Bias

1352

in mean velocities and noise in variances and covariances measured using a multistatic acoustic

1353

profiler: The Nortek Vectrino Profiler. Meas. Sci. Technol., 28, 075302, 25p.

1354

Tinterri, R. (2011) Combined flow sedimentary structures and the genetic link between

1355

sigmoidal- and hummocky-cross stratification. GeoActa, 10, 43-85.

1356

Tinterri, R. and Muzzi Magalhaes, P. (2011) Synsedimentary structural control on foredeep

1357

turbidites related to basin segmentation: facies response to the increase in tectonic confinement

1358

(Marnoso-arenacea Formation, Miocene, Northern Apennines, Italy). Mar. Petrol. Geol., 67,

1359

81-110.

1360

Tinterri, R., Muzzi Magalhaes, P., Tagliaferri, A. and Cunha, R.S. (2016) Convolute

1361

laminations and load structures in turbidites as indicators of flow reflections and decelerations

1362

against bounding slopes. Examples from the Marnoso-arenacea Formation (northern Italy) and

1363

Annot Sandstones (south eastern France). Sed. Geol., 344, 382-407.

1364

Tinterri, R., Mazza, T. and Muzzi Magalhaes, P. (2022) Contained-reflected megaturbidites

1365

of the Marnoso-arenacea Formation (Contessa Key Bed) and Helminthoid Flysches (Northern

1366

Apennines, Italy) and Hecho Group (South-Western Pyrenees). Front. Earth. Sci., 25, 817012.

1367This is a non-peer reviewed preprint submitted to EarthArxiv

Tőkés, L. and Patacci, M. (2018) Quantifying tabularity of turbidite beds and its relationship

1368

to the inferred degrees of basin confinement. Mar. Petrol. Geol., 97, 659-671.

1369

Wei, T., Peakall, J., Parsons, D.R., Chen, Z., Zhao, B. and Best, J.L. (2013) Three-

1370

dimensional gravity-current flow within a subaqueous bend: Spatial evolution and force

1371

balance variations. Sedimentology, 60, 1668–1680.

1372

Wright, L.D., Friedrichs, C.T., and Scully, M.E. (2002) Pulsational gravity-driven sediment

1373

transport on two energetic shelves. Cont. Shelf Res., 22, 2443–2460.

1374

Yalin, M. S. (1971) Theory of hydraulic models. Macmillan, London, United Kingdom, 266pp.

1375

1376

1377

1378

1379

1380

1381

1382

1383

1384

1385This is a non-peer reviewed preprint submitted to EarthArxiv

SUPPLEMENTARY TEXT

1386

Supporting Information 1: Derivation of the input parameters for the estimation of the

1387

Flow Reynolds number and densimetric Froude number

1388

Flow Reynolds number and densimetric Froude number were estimated for the experimental

1389

density current recorded at 3 m downstream from the channel mouth along the channel-basin

1390

centreline in the unconfined reference experiment. They were computed by Equations 1, 3 and

1391

4 (see main text), with input parameters shown in Table 2.

1392

Notably, the overall flow height ℎ (0.11 m) was observed directly from the time-averaged

1393

profiles of downstream velocity (Fig. 3G) at the measurement position, where the downstream

1394

velocity recorded by the UVP reaches zero at the top of the flow. Additionally, two input

1395

parameters were calculated from the time-averaged profiles of downstream velocity and

1396

density (Fig. 3G) at this position: depth-averaged downstream velocity 𝑈, and depth-averaged

1397

density of the current 𝜌𝑠 . They were estimated by averaging the velocity or density values

1398

recorded or extrapolated at regularly spaced height intervals (0.05 m) over the full depth of the

1399

flow, respectively.

1400

1401

SUPPLEMENTARY FIGURES AND TABLES

1402

TABLE S1. Set-up parameters for the Ultrasonic Velocity Profiler (UVP) and Acoustic

1403

Doppler Velocimeter (ADV).

1404

UVP parameters ADV parameters

Instrument Met-Flow UVP Monitor 4 Instrument Vectrino Doppler Velocimeter

Frequency 4 Hz Frequency 100 Hz

Ultrasound speed in water 1480 m s-1 Sound speed in water 1480 m s-1

Number of channels 128 Number of transducers 4

Number of profiles 1000 Range to first cell 0.040 m

Sampling period 11 ms Range to last cell 0.070 m

Axis velocity range 0.256 m s-1 Cell size 0.001 mThis is a non-peer reviewed preprint submitted to EarthArxiv

Minimum axis velocity -0.128 m s-1 Number of cells 31

Maximum axis velocity 0.128 m s-1 Streamwise velocity range 0.300 m s-1

Minimum measurement distance 4.995 mm Horizontal velocity range 1.399 m s-1

Maximum measurement distance 99.715 mm Vertical velocity range 0.372 m s-1

1405

View publication stats