Geology Reference
In-Depth Information
material can be quite high. Part of this description
also applies to other areas where piping and gul-
lied badlands develop on geological substrata of
different ages and materials (e.g. in Tuscany, Italy,
on Plio-Pleistocene marine sediments: Torri &
Bryan, 1997) ). The presence of vegetation, which
favours the formation of large connected pores
due to root growth, also favours leaching
with translocation of sodium. These processes
strongly undermine soil resistance to concentrated
flow erosion. Faulkner
et al
. (2008) used these
observations to explain the coupling of pipes to a
rejuvenating channel in Mocatán. A series of inci-
sions reconnects slopes and channels, causing an
increase in the hydraulic gradient along the slopes.
This in turn favours piping and pipe enlargement,
with subsequent collapse of pipe roofs. Hence lat-
eral gullies can develop quickly. The new (lateral)
gullies have walls along which similar processes
can develop, first piping and then further gullying.
This feedback process will only end after a long
period of slope-base stability. Examples of these
types of processes can be observed in many places
with badlands such as the
biancane
badlands of
Tuscany (Torri & Bryan, 1997).
Radoane
et al
. (1995) reported on an extensive
study of gully head advancement in Moldavia
(Romania) and proposed the following equations:
Burkard and Kostaschuk (1997) examined the
behaviour of 44 gullies cut into clays of glaciola-
custrine and glacial origin along the eastern
shoreline of Lake Huron, Canada, over a 62-year
period (using aerial photographs) and proposed
the following equation:
0.59
R
=
0.3996
A
(19.16)
where
R
is the mean annual gully area growth
(m
2
y
−1
) and
A
is the drainage area (m
2
)
Vandekerckhove
et al
. (2001) monitored 46
active bank gully heads in southeast Spain and
found that the present drainage basin area (
A
, m
2
)
was the most important factor explaining mean
short-term (2 years) gully headcut retreat rate
(
R
, m
3
y
−1
), following the equation:
(19.17)
0.38
R
=
0.04
A
Vandekerckhove
et al
. (2003) showed that when
predicting
R
in southeast Spain, the weight given
to drainage basin area (
A
) increases from the short
term (i.e. a few years) to the long term (decades,
centuries), and attribute this to several reasons:
(1)
So far, most studies have focused on the
medium-term retreat rate of gullies. Little is
known about the processes and factors controlling
the short-term gully head erosion of gullies.
Predicting long-term gully head retreat rates seems
to be easier than short-term retreat rates because
of the stochastic nature of some gully wall sub-
processes such as tension crack development, soil
toppling and soil fall, piping and fluting.
(2)
Spatial rainfall variability may be held respon-
sible for important variations in short-term head-
cut retreat rates within the study areas.
(3)
In the long term, one may expect an increased
contribution of extreme rainfall events to gully
headcut retreat whereby the role of the drainage
basin area becomes more pronounced, as contrib-
uting runoff is produced from a larger fraction of
the gully catchment during such events and run-
off transmission losses are smaller compared with
low-intensity rain events.
Several attempts to model gully growth, based on
theoretical-physical considerations, have been
bc d e
R
=
aA L E P
(for 22 gullies
cut in marls and clays)
(19.14)
R
=+ + + +
a
bA
cE
dL
eP
(19.15)
(for 16 gullies cut in sandy rocks)
where
R
is the medium-term (14 years) mean
retreat rate of the gully head (m y
−1
)
, A
is the
drainage basin area upstream of the gully head
(ha)
, L
is the gully length (m)
, E
is the relief energy
of the drainage basin (m)
, P
is the drainage basin
inclination (m (100 m)
−1
), and
a
,
b
,
c
,
d
and
e
are
empirical coefficients or exponents.
By far the most important independent varia-
ble controlling gully recession rate was drainage-
basin area, explaining 54% of the variance in the
case of marls and clays, and 68% in the case of
sandy rocks.
