Slope systems - Arid Zone Geomorphology - page 214

thus explain variability of response in subsequent events
of the same magnitude.
A final major control on badland formation and evo-
lution is that of human activity. Gullying as a result of
vegetation clearance has been suggested as a triggering
mechanism for the subsequent spread of erosion at times
ranging from prehistoric to the present, over a wide ge-
ographical range (Italy: Torri et al ., 1999; Greece: King
and Sturdy, 1994; France: Ballais, 1996; Spain: Castro
et al. , 1998; South Africa: Boardman et al ., 2003; Easter
Island: Mieth and Bork, 2005; Tanzania: Eriksson, Reuter-
swald and Christiansson, 2003). As noted above, gullying
can have multiple causes, and in some areas more de-
tailed landscape-history reconstructions have shown the
initial phase of erosion may not have been anthropic,
even if it was subsequently exacerbated by human action
(e.g. in North Dakota: Gonzalez, 2001). There has also
been much emphasis on badland reclamation in recent
years in Mediterranean Europe, largely as a consequence
of agricultural subsidies, e.g. using mechanical levelling
(Phillips, 1998a; Clarke and Rendell, 2000; Capolongo
et al ., 2008) or various forms of green engineering (Rey,
2009). Such approaches have not often been successful,
even at fully reestablishing native vegetation to stabilise
slopes at the humid end of the spectrum (Vallauri, Aron-
son and Barbero, 2002), and have often led to significant
increases in surface-erosion rates (Piccarreta et al ., 2006;
Robinson and Phillips, 2001), piping (Romero Dıaz et al .,
2007) and mass movements (Linares et al. , 2002). In sev-
eral cases, opposition to this reclamation has been on the
basis that badlands form a significant part of the aesthetic
and cultural landscape (Phillips, 1998b, 1998c, and the
discussion in Wainwright and Thornes, 2003).

sition in badlands, especially on micropediments and in
the development of alluvial valley bottoms (Figure 10.3).
Gullying has been noted above to be an important pro-
cess in the initiation of badlands, but is also a significant
source for sediment production (Kirkby and Bull, 2000;
Nogueras et al ., 2000). The review of Poesen et al . (2003)
suggests gully erosion may make up anything between 10
and 94 % of total catchment erosion, although none of
their figures are explicitly derived from badlands, so it is
hard to be more specific here. As noted experimentally
by Kosov et al . (1978, cited in Sidorchuk, 1999), during
the first 5 % of the lifetime of a gully, it will typically
have developed >90 % of its length, 85 % of its depth,
60 % of its area and 35 % of its volume (Figure 10.10).
After about 15 % of its lifetime, the length has stopped
extending (usually by running out of a catchment area
sufficient to provide enough concentrated flow: Faulkner,
1974; Thornes, 1985; Kemp nee Marchington, 1990), the
depth is nearly 90 %, the area 85 % and the volume about
55 % of the final amount. These nonlinear patterns of
evolution may go a long way in explaining the apparent
long-term stability of some gullied badland systems, as
noted above. They are also consistent with the observa-
tion that even if the Zin badlands in the Negev have at
times eroded more rapidly, the mean rate of ground low-
ering is equivalent to 0.75 mm/a over a period of 75 ka
(Yair, Goldberg and Brimer, 1982).
Piping is extensively found on clay-rich bedrocks (e.g.
Farifteh and Soeters, 1999; Torri and Bryan, 1997; but
cf. Howard, 2009). Pipes are hydraulically efficient and
erode rapidly, both by mechanical erosion and dissolution
(hence the alternative name of 'pseudokarst': Halliday,
2007). Direct measurements on badland pipe flows are
sparse, but in the Chinese loess belt, Zhu, Luk and Cai
(2002) showed that for 65 % of storm events, pipe flows
preceded the main runoff peak, and even for short storms,
pipe flows could peak up to 30 minutes before the main
subbasin response. They found a wide range of sediment
concentrations in the pipe flows, with a maximum of
893.2 g/L. While some pipes are relatively ephemeral fea-
tures and may form, be destroyed and reform on a storm-
to-storm or seasonal basis - especially beneath 'popcorn'-
type surfaces - larger, structural pipes may be much more
permanent features of badlands. All sizes of pipe are
liable to eventual collapse as progressive erosion leads
to enlargement and thus instability of the pipe roof. Pipe
collapse is thus implicated in rill and gully initiation. This
process is likely to be cyclical, as the new rills and gullies
will lead to a drop in the perched watertable during a
storm event and thus the formation of new pipes following
lower lines of weakness in the bedrock. Pipe development

10.2.1 Processes and rates of
badland evolution

Because of the high perceived rates of badland evolu-
tion, there has been much attention on quantification of
these rates. Comparison of plot- and catchment-based
rates of overland flow is problematic because they are
highly scale-dependent in a nonlinear way (Parsons et al .,
2004, 2006). Point measurements of surface lowering us-
ing erosion pins may be less problematic in this respect,
and show rates ranging from 0.7 mm/a in caprocks to
77 mm/a in slope-foot settings (Table 10.1). However,
caution must still be applied when interpreting these data,
as measurements may tend to be focused where there is
measurable erosion, leading to bias, and some surface
change may be due to surface expansion and contraction,