Generic autogenic behaviour in fluvial systems: lessons from experimental studies - From Depositional Systems to Sedimentary Successions on the Norwegian Continental Margin - page 11

proceeds faster than aggradation of the river so
that the latter can adjust to the formation of a top-
ographic low by the deposition of overbank mate-
rial in the form of splays and sheet sands (see
experiments of Hickson et al . 2005).
Kim & Paola's (2007) experimental studies of
sedimentation in an experimental relay ramp
showed that autogenic cycles developed stratal
packages of subaerial prograding lacustrine delta
deposits bounded by fluvial aggradation units
under constant discharge and sediment yield.
These cycles were formed by strong variations in
sediment delivery associated with tectonically-
driven routing of river flow across and around
the footwall uplift. Flow patterns of sheet flow
and channelised flow ('avulsion cycles') became
five times longer during the active subsidence
(delayed the backfilling process). The period of
the tectonic-driven autogenic processes was
inferred to be of the order of 10 kyr to 100 kyr,
which would be much lower than the normal
autogenic behaviour.
Hence, the response in aggradation rate to tec-
tonic change varies strongly with the kind of kin-
ematics. Active fault scarps could make a fluvial
stretch to subside instantly, bringing the system
from fill-up to start-up stage, herewith increasing
aggradation rates instantly. Basinward tilting of
the fluvial profile as occurs, for instance, in pas-
sive margin settings would decrease aggradation
rates, because the profile is tilted towards its
grade.

However, the experimental research mentioned
here give reasons toward a more positive attitude.
In spite of the fact that the experiments are not
scaled hydraulically, the experimentalist has the
great advantage of looking at a natural 'forward'
model with similarity of process where the prod-
uct can be studied in relation to input conditions
(Paola, 2000; Paola et al . 2009).
The existence of scale-invariable morphological
features like channels, bars and lobes hints to the
similarity of process that is obtained in laboratory
models. Sediment transport averaged over suffi-
ciently long time periods can be predicted by
diffusion (Paola et al . 1992). The crude fluvial
architecture stemming from aggradation as well as
from variations in depositional slope characteris-
tic for the various river types can be simulated
easily by using different exponents in a non-linear
diffusion equation (Postma et al . 2008). Fig. 6
shows a dimensionless plot of aggradation rate by
normalised sediment yield q in /q out against time (T)
relative to the timescale that the fluvial system
requires to reach grade (T eq ). The equilibrium
timescale is the ratio of L 2 /k, with L being a
length scale which is given by the river's active
depositional trajectory and k the diffusivity coef-
ficient, which is related to the discharge (Paola
et al ., 1992). The active depositional trajectory rel-
evant for autogenic behaviour (avulsion) would be
the backfill trajectory. With mean diffusivities of
the order of 0.01 km 2 /yr (Paola et al ., 1992), chan-
nel depth of 7 m to 10 m and slopes of the order of
0.0001 (from Kleinhans et al ., 2008), most low gra-
dient rivers in the delta plain have a backwater
length of approximately 25 km, so that T eq for the
reach is about 60 kyr. For low gradient rivers a lin-
ear diffusion equation for simulation of sediment
transport over long time intervals is justified (e.g.
Paola et al ., 1992), so that the start-up stage is
almost non-existent (Fig. 6). However, it should be
noted that both the length scale and the diffusivity
coefficient vary dynamically and with that the
calculated equilibrium time. Hence, its value
should be treated with caution and only in a first
order of approach.
Allogenic forcing brings the system continu-
ously out of balance and changes its accumulation
space and herewith the aggradation rate, as was
discussed above. In asking 'is it possible to predict
the change in aggradation rate?' the author believes
it is possible to predict the change in a first order
of approach. If the time period for the change
in accumulation space is much faster than T eq ,

DISCUSSION

In a discussion about how well fluvial architec-
ture can be predicted in surface and subsurface
analyses, Miall (2006) concluded that little can be
expected beyond the provision of a general start-
ing point. He argued that the variety of fluvial
forms in modern rivers and the ancient record is
vast, making the choice of an appropriate ana-
logue very difficult. Fluvial style varies laterally
or vertically through most real stratigraphic units
because of the constant interplay of several allo-
genic controls acting on different time scales.
Given the complex-response character of fluvial
systems to allogenic forcing and including the
tendency for systems to lag behind changes in
forcing functions at varying rates, the predictabil-
ity of fluvial architecture aerially and stratigraphi-
cally must be considered quite limited.