Geology Reference
In-Depth Information
6
Scaling Soil Erosion Models
in Space and Time
R.E. BRAZIER
1
, C.J. HUTTON
1
, A.J. PARSONS
2
AND J. WAINWRIGHT
2
1
School of Geography, University of Exeter, Exeter, UK
2
Department of Geography, University of Sheffield, Sheffield, United Kingdom
6.1 Introduction
(e.g. Rendell, 1982), or more recently, direct photo-
grammetry or laser scanning of the surface (e.g.
Chandler, 1999). The resulting measurements are
thus in units of height/depth of ground gained or
lost per unit time [L T
−1
]. The second - far more
commonly used - employs the direct capture or
sampling of sediment passing a point in the land-
scape. Whereas the first set of techniques comprise
point measurements, the second are spatially aver-
aged, albeit not in a straightforward way as will be
seen. Direct capture may employ artificial stores
(e.g. Gerlach, 1967) or reservoirs and lakes
(McManus & Duck, 1985; Rowan
et al
., 1995;
Verstraeten & Poesen, 2000), while sampling may
use slot- or wheel-based designs (Brakensiek
et al
.,
1979) or, increasingly, pump samplers originally
designed for water samples (Bilotta
et al
., 2008).
A range of different measurement units is typically
produced. In artificial sediment traps, the weight of
sediment per unit time [M T
−1
] is used, while reser-
voir or lake studies yield net sedimentation rates
[L T
−1
]. Depending on the method of sediment
capture, slot- or wheel-based samplers may produce
weights per unit time, or combined water and sedi-
ment samples, as with the pump sampling approach.
In the latter, it is common to calculate a sediment
concentration [M L
−3
], even though this is based on
a direct measurement of the sediment weight, and
despite the fact that the suspension is an artificial
consequence of the sampling technique. Compar-
ison of these different methodologies shows that
quite different measurements are being made, and a
number of assumptions are generally necessary to
Consider a soil particle sitting on a slope. During
a storm event and its immediate aftermath, the
particle is subjected to a range of forces that act to
detach and transport it both within the slope and
ultimately through the catchment in which it is
located. These forces will cause the particle to
move at a range of velocities and for a range of
distances, although for the majority of the time
the forces will be insufficient to exceed thresh-
olds for movement and the particle will remain
stationary. Soil erosion - detachment, transport
and deposition - is made up of very large numbers
of these irregular stationary periods and steps.
Characterization and prediction of erosion is thus
a matter of integrating these very large numbers
of individual movements to produce fluxes of
sediment movement. Were erosion processes and
their direct controls constant, in space and time,
the central limit theorem should make this task
very simple. Unfortunately they are not, and the
problems in characterizing erosion rates are fur-
ther compounded by spatial and temporal varia-
bility in slope and catchment characteristics.
Direct measurement of erosion rates employs
two broad techniques. The first considers the
change in surface elevation directly using methods
such as microtopography meters/erosion bridges
