Generic autogenic behaviour in fluvial systems: lessons from experimental studies - From Depositional Systems to Sedimentary Successions on the Norwegian Continental Margin - page 5

Stouthamer & Berendsen, 2001). Peat formation
potentially influences avulsion by inhibiting lat-
eral migration and increasing aggradation in the
channel belt. Peat compaction and oxidation in
flood basins also leads to relief amplification and
to super-elevation of channel belts (Van Asselen
et al . 2009).
Cyclic avulsion processes in the Yellow River
delta were forced by rapid delta progradation
causing the river to adjust its channel belt profile
by aggradation (Kriele et al ., 1998). At some point
the aggradation led to an increase of transverse
slopes causing the channel to avulse in another
direction. It is noted here that this process is in
contrast with the initiation of avulsion on stream-
dominated alluvial fans as determined from our
experimental studies, where sheet flow increases
the gradient of the entire apex to levels of instabil-
ity and new channel incision. In braid plains, gra-
dient advantage in bifurcations does play a similar
role, yet occurs on much smaller temporal and
spatial scales.
In summary, autogenic behaviour in alluvial
fans and braided river systems is different from
that in moderate-gradient and low-gradient river
systems and is most strikingly different by the
effect that backwater has on the channel belt
aggradation. The effect of backwater is virtually
lacking in the moderate and steeply graded sys-
tems (see equation 1), while the water flow in
the channels is often close to supercritical
(Sheets et al . 2002; CGER, 1996; Hoyal & Sheets,
2009); so caution is needed when applying
experimental studies of those systems to low
gradient rivers. However, in all cases backfilling
of the channels is a prerequisite for avulsion,
since backfilling elevates the channel above its
surroundings. In the case of alluvial fan systems,
the avulsion trigger is clearly related to steepen-
ing of the apex by the sheet flow, which can only
commence if backfilling is completed. In the
case of braided river systems, the trigger is by a
growing advantage of one bifurcate over the
other. The avulsion process of the braided river
is thus, in this respect, similar to that of the sin-
gle river, where avulsion also starts with a bifur-
cation but where the change from bifurcation to
avulsion is up to three orders of magnitude
slower. Avulsion frequencies vary greatly among
modern river systems, with a lowest rate of 28
years for the Kosi River in India and up to 1400
years for the Mississippi River (Slingerland &
Smith, 2004).

FREQUENCY OF AUTOGENIC
PROCESSES

The rate of backfilling (aggradation in the channel
belt) defines the frequency of autogenic processes
(cf. Van Dijk et al . 2009). Backfilling commences
where there is sufficient reduction in channel
slope to force deposition. If the lower boundary is
the shoreline, then reduction of slope is obtained
by progradation of the shoreline and the creation
of a mouth bar (e.g. Kriele et al . 1998). If the flu-
vial system is prograding over a (flood) plain, it
must also be the reduction in slope by prograda-
tion of the system that forces deposition and
formation of a mid-channel bar, heralding the
backfilling. The accumulation space and the rate
at which the required sediment volume aggrades
the channel both determine the avulsion fre-
quency. Bryant et al . (1995), on the basis of their
experimental results, were the first to link avul-
sion rate with aggradation rate, although they did
not measure the aggradation but simply took
sediment yield at the apex of the fan as a proxy for
aggradation rate. Here, it is important to realise
that it is not the total of the supplied sediment is
important, but how much of the supply is used for
aggradation of the bed. Supplied sediment that is
not used for aggradation bypasses the channel and
is used for progradation of the system at the chan-
nel mouth (telescoping fans). If the rate of backfill-
ing is slow, observations from experiments (Van
Dijk et al . 2009) indicate that the river valley has
time to deepen and widen, which increases its
accumulation space causing a negative feedback
to avulsion frequency, thus reducing it. Incipient
relief, channel length and fan-produced local
surface irregularities (lobes, scours, channels
and bars) are most probably causes for observed
deviations in cycle duration and the timing of
re-incisions.
Van Dijk et al . (2012) showed how frequency
of autogenic processes in their steep gradient
systems is related to channel backfill rate. The
frequency in their experimental alluvial fans,
which prograded over a near horizontal plain,
appeared to be much higher than for fan deltas.
Since the upper boundary conditions for the
alluvial fans were the same in both experimental
set ups, the change in frequency must have been
caused by the only difference between the exper-
imental fans, i.e. the presence of a shoreline.
They demonstrated that the presence of the
shoreline caused different aggradation rates on