Geology Reference
In-Depth Information
In some industries the timescales required of con-
tainment are long. For instance, worldwide it is
generally required that even the most benign
waste from the nuclear industry must be con-
tained for a minimum of 1000 years (e.g. Willgoose
& Riley, 1998a,b; Crowell
et al
., 2005). Long
design lifetimes are also being increasingly
required worldwide for the rehabilitation of mine
sites. This is particularly true when: (1) it is
anticipated there will be no active management
of the site post-closure; (2) governments perceive
that they may be required to fund clean-up in the
event of failure; or (3) governments may be held
legally responsible in the event of failure because
they signed off the containment strategy.
Often this waste has to be contained above the
ground surface. In this case, erosion of the encap-
sulation structure is a primary failure mode.
Often above-ground structures are difficult to
avoid. For instance, in the mining industry when
rock is mined it is fragmented as part of the min-
ing operation. The fragmentation of the rock
increases its volume by 30-40%, primarily
because of the newly created air voids between
the rock fragments relative to what was previ-
ously solid rock. This means that for mining
operations where the metal being extracted is
only a small percentage of the volume (typically
all mines except coal, iron ore, and aluminium) it
is not possible to place all of the mine waste back
in the hole from which it was mined, so some of
the waste needs to be stored above the ground.
Even for the exceptions there can be operational
reasons why placing the waste back in the hole is
unfeasible or very difficult.
The first application of a landform evolution
model to the assessment of the long-term stabil-
ity of an encapsulation structure was by Willgoose
and Riley (1993, 1998a,b) who examined an above-
ground structure (proposed as part of the rehabili-
tation design) for the Ranger Uranium Mine,
Northern Territory, Australia. The discussion
below follows Willgoose and Riley (1998a,b) with
updates reflecting follow-on work.
Field runoff and erosion plots (with a range of
areas and slopes) were used to calibrate an event
hydrology model and an event erosion model.
The event hydrology model was then used to
generate a 20-year, 15-minute resolution runoff
series using recorded pluviograph data from a
nearby meteorological station. This runoff series
was then used with the event erosion model to
generate a 15-minute resolution erosion series.
The long-term average erosion model in the
SIBERIA LEM, which relates sediment transport
rate to catchment area and slope, was then cali-
brated to this erosion series to yield the transport
law to be used in the long-term simulations. This
calibration process meant that the erosion law
used for the landform evolution simulations
yielded the result of the average sediment trans-
port, and did not model the effect of individual
runoff and erosion events. Rainsplash erosion
effects were calibrated by comparing paired plots
that were covered with shadecloth and plots
without rainsplash protection. Soil creep and
other mass movement processes were not mod-
elled because they were not believed to be impor-
tant at this site.
The initial landform and the landform after
1000 years of erosion are shown in Fig. 18.1. The
first thing to note is that the erosion that has
occurred is not uniform in space. There are
regions of localized high erosion (i.e. the valleys)
separated by regions of low erosion (i.e. the
ridges). The high-erosion regions have total ero-
sion depths of up to 8 m, while some of the ridges
have barely eroded. Note that due to the vertical
exaggeration in the figures, the high-erosion areas
look like gullies but are in fact more like valleys
(8 m deep by 120 m wide). The valleys are initi-
ated at the transition from the low slope top of
the structure to the higher slope batters. They
subsequently propagate upstream from the tran-
sition, while simultaneously cutting down. On
the flatter natural areas surrounding the landform
(and downstream of the valleys in the structure),
alluvial fans are created. Over time they increase
in depth and extent from the structure. These
fans are barely visible in Fig. 18.1. The rapid val-
ley downcutting stops when the valley floor cuts
down to the height of the depositing alluvial fan
(Fig. 18.2). Subsequently, depending on the geom-
etry of the landform upstream of the initiation
