Geoscience Reference
In-Depth Information
Fig. 1 Schematic drawings of experimental facilities: (a) viscous-fluid flume (after Coleman and
Eling
2000
) and (b) water tunnel (after Coleman et al.
2003
)
At the start of a bed development run, the bed was smoothed, the settings for the
desired flow were established, and then the flow was initiated. For the first runs,
bedforms were simply observed. For the later runs, the flow was halted at selected
stages of bed development and the centreline bed profile was measured with the
channel remaining filled with oil. The bed profiles were measured using a bed-
profile measurement system utilising an ultrasonic depthsounder (Coleman
1997
),
this system being configured to record bed elevation measurements to within
0.4 mm every 1.2 mm along the flume. The centreline velocity profile midway
along the test section was measured using a laser Doppler velocimeter for each run
of measured bed profiles.
The results of the tests undertaken indicate that both antidunes and also the seed
waves leading to dunes and ripples (Fig.
2
) can be generated from plane-bed
conditions in open-channel laminar flow, the lengths, shapes, and patterns of gen-
eration and growth (Fig.
3
) for these sand waves in laminar flows being consistent
with observations for alluvial flows (Coleman and Melville
1994
; Coleman et al.
1998
; Coleman and Eling
2000
). In particular, the seed waves are of a preferred
wavelength that is relatively insensitive to the characteristics of the applied flow
and primarily a function of the size of the sediment. Coleman and Eling (
2000
)
propose that these seed-wave lengths
for alluvial and laminar open-channel flows
over beds of quartz and lightweight sediments of size
d
l
0.2 to
d
1.6 mm can
¼
¼
175
d
0.75
, where
be simply described by
and
d
are expressed in millimetres.
With both ripples and dunes being postulated to subsequently develop from seed
l
¼
l
