MassFLOW-3DTM as a simulation tool for turbidity currents: some preliminary results - From Depositional Systems to Sedimentary Successions on the Norwegian Continental Margin - page 599

in sinuous submarine channels. The final case
study reports the preliminary results of a study of
the flow structure of turbidity currents in a natural-
scale sinuous channel in an attempt to contribute to
the understanding of these heterogeneous systems.

was roughened with medium sand (ø = 0.5 mm).
The geometry and dimensions of the flow inlet
were identical to the cross-sectional geometry of
the channel. Channel slope was set to 1°. Simulation
domain, containing the channel surface, consisted
of a Cartesian orthogonal mesh with a total number
of 784,259 active cells. The mesh extended 265 m
above the channel floor and had a pressure-
specified top boundary to account for hydrostatic
pressure at a water depth of 1 km. The resolution of
the mesh was uniform up to 20 m above the channel
floor and then gradually decreased towards the top.
The smallest and largest cell had the dimensions of
14.8 m × 14.8 m × 1.5 m and 14.7 m × 14.7 m × 11 m
(XYZ format), respectively.
Eight simulations runs were performed in total.
In each run the turbidity current was released from
the inlet for a duration of 1 hr and consisted of two
separate grain sizes; medium sand (ø = 0.5 mm)
and coarse sand (ø = 1.0 mm) with a density of
2648 kg m -3 . The medium/coarse sand ratio was
kept at 4:1 during all the runs, independent of

Numerical simulation set-up

Turbidity currents were released into a non-
erodible pre-defined channel consisting of an
initial 1 km-long straight segment and subsequent
three channel bends out of which two were 90° and
one 180° (Fig. 12A). The study focused on the 2nd,
a 180° bend, which had a radius of 600 m and a
bend wavelength of 1800 m. The relatively long
straight segment preceding the bends was intro-
duced to stabilise the vertical velocity and density
structure of the flow, so that the incoming flow at
the domain boundary had a uniform vertical struc-
ture. Channel width and depth was set to 150 m
and 10 m, respectively. Channel banks were con-
structed at the angle of repose (30°) and the surface

(A)

(B)

flow inlet 1.0

0.8

channel slope=1°

0.6

Simulation study
This study
Corney et al. (2006)

0.4

Imran et al. (2007)
Experimental study

0.2 Corney et al. (2006)

Straub et al. (2008)
Sumner et al. (2008)
Keevil et al. (2006)
Imran et al. (2007)

0.0

Channel width=150m
Channel depth=10m

−0.2 0.0 0.2 0.4 0.6 0.8 1.0

Downstream velocity (U/U max )

(C) (D)

Total cell count = 1204310
Smallest cell = 14.8 × 14.8 × 1.5 m (XYZ format)

Total cell count = 450182
Smallest cell = 19.5 × 19.5 × 2.3 m (XYZ format)

ve × 3
25 m

ve ×3
25 m

1025 1256

1025 1186

flow density (kg/m 3 )

flow density (kg/m 3 )

Fig. 12. (A) Dimensions and terminology of the pre-defined non-erodible channel surface used for simulations. (B) Plot
showing the comparison of previously published downstream-velocity profiles with profile simulated in the present study.
(C, D) Domain profiles demonstrating the independence of the simulated output from the resolution of the simulation grid.
In D, the cell count is less than a half of the grid shown in C; however, there is virtually no difference in the calculated
velocity and density (concentration) magnitude.