Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 43 No.

1 March 2012 ISSN 0046-5828

Some Mining Applications of Unsaturated Soil Mechanics

D.J. Williams1
1
School of Civil Engineering, The University of Queensland, Brisbane, Australia
E-mail: [email protected]

ABSTRACT: Unsaturated soil mechanics continues to play poor relation to saturated soil mechanics, although an unsaturated soil at a given
density is stronger, less compressible and less permeable (i.e. performs better) than the same soil in a saturated state. There are many
examples of unsaturated conditions in the mining field, including the wetting-up and drain-down of initially dry surface waste rock dumps;
the irrigation and drain-down of heap leach materials; the drain-down, desiccation and rewetting of mine tailings; the dewatering of mineral
products such as coal; the shear strength and compressibility of stored mine wastes; and the performance of geo-covers placed on mine
wastes on rehabilitation. This paper highlights the key unsaturated soil mechanics parameters, overviews the nature of mining and
processing wastes, and some products, and discusses the issues involved. Some applications of unsaturated soil mechanics addressing the
shear strength, compressibility and permeability of mine wastes, and mineral products, are presented, together with data to highlight them.

1. INTRODUCTION The exploitation of mineral ore bodies involves either open pit or
underground mining, followed by mineral processing. Open pit
The key parameters of interest in soil mechanics are: (i) the strength, mining produces coarse-grained wastes including overburden or
or the capacity of a soil to support load; (ii) the compressibility, or waste rock, while underground mining produces only limited
the deformation of a soil under an applied load; and (iii) the coarse-grained wastes from the excavation of access ways and
permeability (hydraulic conductivity), or the rate at which a soil will ventilation shafts.
drain and deform under an applied load. These apply to both Mineral processing of crushed and ground ores from open pit or
(water) saturated and unsaturated conditions. underground mining operations produces fine-grained wastes known
The shear strength of saturated soils is typically assessed by in as tailings. The smelting of metalliferous concentrates typically
situ vane shear strength testing of soft cohesive soils, or by produces coarse-grained slag or scats. The washing of run-of-mine
laboratory direct shear or triaxial shear strength testing of intact coal produces both coarse reject and tailings.
stiffer soils. The compressibility of saturated cohesive soils is Low grade ore bodies that do not warrant the high cost of
typically assessed by laboratory consolidation testing. The saturated grinding may be leached in a heap to recover the commodity. The
hydraulic conductivity of soils is assessed in the laboratory by spent heap leach material constitutes another form of coarse-grained
constant or falling head permeameter testing, or indirectly from the waste, which will be contaminated with the process chemicals
results of consolidometer testing, depending on the permeability of applied on leaching.
the soil. The mining and processing of mineral ores has the potential to
Estimates of the shear strength and compressibility of generate contaminated water, which may impact the surrounding
unsaturated soils can be obtained by in situ vane shear strength environment, transported by surface or ground waters, or by the
testing of desiccated soft soil profiles, or by laboratory direct shear wind.
or triaxial testing of soils from their in situ unsaturated state.
Estimates of the hydraulic conductivity of an unsaturated soil can be
2.2 Hard Rock Metalliferous Ore Body Wastes
obtained by a combination of its soil water characteristic curve
(SWCC) and its saturated hydraulic conductivity. The shape of the Open pit mining of hard rock, disseminated, metalliferous ore
unsaturated hydraulic conductivity relationship with matric suction bodies, such as copper, gold, nickel, and zinc, involves blasting of
is a function of the shape of the soil’s SWCC, and the saturated the rock to produce particles typically finer than about 1 m in size.
hydraulic conductivity sets the intercept at near-zero matric suction The barren or unmineralised waste rock, which may comprise about
or near-saturated conditions. half of the rock mined, is typically hauled to a surface waste rock
Simplistically, the shear strength of unsaturated soils can be dump where it is typically end-dumped from a tip-head to form
accounted for by adding a cohesion term for the additional shear angle of repose slopes (see Figure 1).
strength provided by matric suction, which increases the effective For open pit mining, the minimum economic copper ore grade is
stress. It can also be accounted for, simplistically, by laboratory of the order of 1%, while the minimum economic gold ore grade is
testing of soils in their unsaturated state. The greater stiffness of of the order of 1 part per million (1 ppm). The minimum economic
unsaturated soils can simplistically be accounted for by testing the grade of a metalliferous ore will also depend on the stripping ratio of
compressibility of soils in the laboratory in their unsaturated state. waste to ore, and whether more than one metal is present in the ore,
There are also more sophisticated laboratory test apparatuses which could add value and make sub-economic primary ore bodies
available that allow testing under constant matric suction, although viable.
these are not discussed in this paper. It could be argued that in Underground mining of hard rock metalliferous ore bodies
reality the matric suction a soil experiences will change as it is involves blasting of the accessed ore to produce particles typically
loaded and compressed, as will its degree of saturation and finer than a few hundred mm in size. Given its greater cost,
hydraulic conductivity. underground mining is typically only economic for relatively high
These simplistic approaches to determining the key unsaturated grade ore bodies (say up to 10% for copper ore or up to 10 ppm for
parameters for application to predicting the behaviour of unsaturated gold ore). For high grade ores found in veins, very selective
mining and mineral processing wastes, and mineral products, are the extraction may be economic. Open stoping may be economic for
theme of this paper. medium grade ores, while block caving may be economic for more
low grade ores, approaching the cut-off grades for open pit mining.
2. MINING AND MINERAL PROCESSING WASTES The blasted ore is crushed to a maximum particle size of the
order of 15 mm, before being ground in a rod or ball mill to a
2.1 Wastes Produced by Mining and Mineral Processing maximum particle size of the order of 1 mm, to expose the
commodity. The ground ore is then processed to yield the mineral

83
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 43 No.1 March 2012 ISSN 0046-5828

commodity. Typically, mineral processing involves dissolution of and bauxite. The mining of these bulk commodities is typically by
the exposed commodity; e.g. using an acid for metalliferous ores, open pit methods, although as shallow coal deposits are exploited
and cyanide in an alkaline bath (at a pH of about 12 to prevent the there is a move to underground highwall and longwall mining. The
formation of cyanide gas) for gold. The presence of more than one mining of bulk ore bodies generates soil and soft rock spoil, while
metal in the ore can make the ore body more economic, but it can the washing of these ores for beneficiation purposes generates soil-
also complicate processing, potentially making the ore body less like tailings.
economic.
The washing of iron ore, when needed to meet market
specifications, typically produces relatively small volumes of
relatively coarse-grained tailings (see Figure 3).

Figure 1 Typical surface waste rock dump

at a hard rock open pit mine

The waste ground rock flour that remains after the extraction of Figure 3 Typical surface tailings storage facility at an iron ore mine
the commodity (or tailings) are typically disposed as a slurry by
pumping to a surface tailings storage facility (see Figure 2). For The washing of run-of-mine black coal, typically required to
metalliferous ores, the bulk of the crushed and ground ore ends up meet market specifications, produces up to about 25% washery
as tailings. Since the tailings are discharged to the storage facility as wastes, comprising coarse reject (typically -50 mm) and tailings
a slurry, considerable water is also discharged, much greater than (typically -0.5 mm). Conventionally, the coarse reject is disposed of
rainfall in a dry climate, resulting in inevitable seepage from the in surface dumps (see Figure 4), and the tailings in surface tailings
facility during its operation. storage facilities (see Figure 5), although many coal mines in
Australia and Indonesia co-dispose of the washery wastes by
combined pumping to a storage facility (Figure 6). As open pits
become available and mining moves underground, the completed
pits are often used to store coarse reject and tailings, or co-disposed
washery wastes (Figure 7).

Figure 2 Typical surface tailings storage facility

at a hard rock open pit mine

Depending on the geochemistry of the ore and the process water

used, and the process chemicals used, the chemistry of the pore
water within the tailings will vary, dictating the chemistry of any
seepage reporting from the tailings storage facility to the Figure 4 Typical surface coarse reject dump at a black coal mine
environment. In the arid Kalgoorlie mining region of Western
Australia, the shortage of fresh water supplies has forced the use of The washing of bauxite ore generates benign, soil-like tailings
hypersaline groundwater for processing, leading to hypersaline that typically comprise up to about 50% of the raw ore. The washed
tailings and seepage. bauxite is processed under alkaline, high temperature and pressure
conditions to produce alumina, generating alkaline red mud (see
2.3 Soft Rock Bulk Ore Body Wastes Figure 8). The red mud typically comprises up to about 50% of the
process feed. In some cases, the alkaline red mud is neutralised
Soft rock bulk ore bodies include iron ore, steaming (brown or black using sea water prior to disposal.
coal for power generation) and black coking coal (for steel-making),

84
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 43 No.1 March 2012 ISSN 0046-5828

located below the groundwater table, and hence unoxidised). This is

due to the fact that hard rock nickel ores are relatively
straightforward, if energy-intensive, to mine, requiring blasting to
fragment the rock, crushing and grinding to expose the disseminated
commodity; and relatively straightforward to process; this simply
requires a large amount of energy. Nickel laterites are a clayey silt-
sized soil dominates by halocites, which are difficult to handle on
mining, and require sophisticated processing under high temperature
and pressure conditions, generating problematic fine-grained
tailings in a high-density, magnesium-dominated liquor.

Figure 5 Typical surface tailings storage facility

at a black coal mine

Figure 8 Typical surface red mud storage facility

Figure 6 Typical co-disposed washery waste storage

at a black coal mine

Figure 9 Typical tailings disposal following

sand mining by dredging

Other ore bodies extensively exploited include quarried rock,

limestone for cement and lime manufacture, phosphate and potash
for fertiliser production, pyrite for sulfuric acid production, plus a
range of other metals; uranium, diamonds, and so on. These share
some of the waste generation and disposal, and potential
contamination, issues of the other ore bodies referred to above, to
lesser or greater degrees.
Figure 7 Typical in-pit washery waste storage (tailings behind a
2.5 Mine-Impacted Water
coarse reject bund) at a black coal mine
Even with the best efforts to divert clean surface and underground
2.4 Wastes from Some Other Ore Bodies waters around impacted areas, the mining and processing of mineral
ores inevitably generates mine-impacted water. In most
Other ore bodies include mineral sands and nickel laterites. The jurisdictions, mine-impacted water cannot be released to the
mining of mineral sands, of which Australia has about half the environment unless it can be shown to not impact water quality (pH,
world’s reserves, is generally carried out by dredging, and produces total dissolved salts, and turbidity). The consequent storage of
both sand and slimes tailings, with the commodity comprising up to mine-impacted water in surface storages leads to evaporation and
10% of the ore mined. The tailings are re-deposited into the lake the concentration of salts and other contaminants. These can
formed by dredging the mineral sands (see Figure 9). potentially be released to the surrounding environment through
About two-thirds of the world’s nickel reserves are in nickel seepage or overtopping during flooding.
laterites. However, to date the majority of the mined nickel is from
hard rock ore bodies, which are generally nickel sulfides (being
85
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 43 No.1 March 2012 ISSN 0046-5828

The exposure on mining and mineral processing of sulfide-rich until its hydraulic conductivity drops sufficiently that it again holds
mine wastes to the atmosphere, previously unoxidised due to their water in storage.
being buried below the groundwater table in which dissolved
oxygen levels are very low, leads to the generation of oxidation
products. These, when combined with water, generate acid and
metalliferous drainage (AMD), which has the potential to reach and
impact the surrounding environment through surface and ground
water flows. A lowering of the pH leads to the dissolution of metals
present in the wastes, leading to AMD of high acidity and high
dissolved metal concentrations. Under certain circumstances, high
dissolved metal concentrations can occur under near-neutral pH
conditions.
The oxidation of sulfidic waste rock and coarse-grained
processing wastes stored in surface dumps is largely unimpeded, due
to the ready availability of air and rainfall infiltration. In fact, such
dumps, which are typically constructed by end-dumping from a tip-
head, are oxidation reactors, with air very readily supplied through
convection via the coarse-grained base rubble zone formed due to
the natural ravelling of coarser-grained particles to the base of the
dump. The air then passes up the semi-continuous coarser-grained
angle of repose layers, and from there diffuses into the adjacent Figure 10 Progressive fingering observed in a 1/15th-scale model of
finer-grained layers, which present a far higher surface area per unit a 15 m high waste rock dump
volume available for oxidation. Infiltrating rainfall and bacteria
catalyse the oxidation reactions, and water in excess of what can be The heap leaching of a low grade metalliferous ore is typically
held in storage within the dump emerges at the base of the dump as somewhat of a trial-and-error approach. Heap leach material is
AMD. The base seepage will either percolate into the foundation or irrigated with a leach solution to dissolve the commodity, which is
emerge from the toe of the dump from topographic low points. collected as seepage at the base of the pad for further processing.
As for the wetting-up of a waste dump by rainfall infiltration, the
3. SOME MINING-RELATED UNSATURATED SOILS wetting-up of a heap leach pad due to irrigation will develop
ISSUES through fingering along preferred pathways. The presence of leach
Some mining-related unsaturated soils issues include: (i) the rainfall solution will tend to enlarge these preferred pathways, further
infiltration-induced wetting-up and drain-down of waste rock and concentrating the flow, and limiting the leaching of finer-grained
coarse-grained processing waste dumps, and the irrigation and materials which expose a far greater surface area of commodity per
drain-down of heap leach materials, (ii) the drain-down, desiccation, unit volume. Hence, heap leaching can be relatively inefficient,
and re-wetting by fresh tailings deposition and rainfall, of stored even ineffective, in maximising the metal yield.
tailings, (iii) the dewatering and atmospheric drying and re-wetting
of product coal, (iv) the bearing capacity and deformation of stored 3.2 Drain-down, Desiccation and Re-Wetting of Tailings
mine wastes, and (v) the performance of mine waste geo-covers. Tailings are typically disposed of as an aqueous slurry by centrifugal
pumping. The slurry concentration will vary with the physical and
3.1 Wetting-Up and Drain-down of Coarse-Grained Wastes chemical nature of the tailings, including the specific gravity of the
Coarse-grained waste rock, spoil, and processing wastes stored in solids. The pumpable solids concentration typically ranges from a
surface dumps act as a “sponge” with respect to rainfall infiltration, low of about 25% solids (percentile mass of solids/mass of solids
due to their relatively porous and permeable surface (Williams, and water) for clay-rich coal mine tailings (low specific gravity), to
2006). This is in contrast to natural hardpan surfaces which, in a dry a high of about 65% for ground hard rock tailings of high specific
climate, allow very little rainfall infiltration. gravity (typically due to them being pyrite-rich). These % solids are
Rainfall infiltration into initially dry coarse-grained waste equivalent to gravimetric moisture contents (percentile mass of
dumps goes largely into storage within the dump. Wetting-up water/mass of solids) of 300 to 54%, and total moisture contents
occurs by progressive fingering (see Figure 10, which shows a (percentile mass of water/mass of solids and water, or 100 - %
1/15th model of a 15 m high waste rock dump, with a finer-grained solids; which can be removed by a combination of mechanical
top layer to simulate a haul truck trafficked surface). dewatering and thermal drying) of 75 to 35%.
Eventually, the dump wets-up sufficiently, but without fully On disposal, the tailings undergo beaching, hydraulic sorting
saturating, that its hydraulic conductivity equals the rate of rainfall (according to particle size and specific gravity), sedimentation and
infiltration, allowing seepage from the base at the same rate as consolidation (accompanied by drainage), and the released
rainfall infiltrates the top, or “continuum breakthrough”. The time supernatant water forms a pond at the downstream end of the beach,
required to reach this point will be a function of the height (available while seepage will also likely be generated. The beaching processes
storage) of the dump, the rainfall and intensity, and the physical and can increase the % solids to between 50% and 70% (gravimetric
geochemical nature of the wastes. The higher the dump, the lower moisture content to between 100% and 43%). The exposed beach
the rainfall and intensity, and the more well-graded the wastes, the then undergoes desiccation, and further consolidation, as the tailings
longer it will take to reach continuum breakthrough. develop matric suction, with the % solids increasing to perhaps
Spreadsheet-based simulations of the wetting-up of waste rock between 60% and 80% (gravimetric moisture content to between
dumps by Williams (2006) found that durable, fresh waste rock will 67% and 25%).
only need to wet-up about 25% of its porosity to achieve continuum On the deposition of fresh tailings or incident heavy rainfall, the
breakthrough, while well-graded, weathered waste rock will need to desiccated tailings will re-wet, possibly resaturating fully, with some
wet-up about 60% of its porosity to achieve this, both corresponding minor decrease in the % solids. On closure, the stored tailings will
to a hydraulic conductivity of about 1.5 x 10-8 m/s. Once continuum again desiccate, and seepage potential will be much reduced.
breakthrough is reached, the dump will drain-down and desaturate

86
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 43 No.1 March 2012 ISSN 0046-5828

3.3 Dewatering and Atmospheric Effects on Product Coal content (%)

% of total
Following beneficiation by washing, the various product coal size 35 20 45
moisture
fractions are subjected to mechanical dewatering, typically by
centrifuging and vacuum filtration, to meet the specified moisture
The trend over time is for Australian product coals to become
content for sale (typically 10 or 11% total moisture content;
Williams, 2009). The product coal size fractions are coarse finer-grained due to improvements in the performance of ultra-fine
(typically 5 to 50 mm), fine (typically 0.2 to 5 mm), and ultra-fine coal recovery methods, the mining of poorer quality coal deposits,
and increased rehandling the raw coal. Hence, the moisture content
(typically up to 0.2 mm; see Figure 11).
of the composite product coal will tend to increase.
Following mechanical dewatering, the re-combined composite
product coal size fractions are stockpiled prior to use or transport to
the port for export. On re-combination, the composite product coal
is unsaturated. On the stockpile, the product coal undergoes
drainage under gravity, cycles of atmospheric drying by the sun and
wind action, and wetting by incident rainfall, plus wetting-up due to
the spraying of water on the stockpile to suppress dust (Williams,
(a) (b) (c) 2010).
Figure 11 Photographs of typical (a) coarse, (b) fine, and (c) ultra-
fine product coal size fractions at their ex-plant moisture states 3.4 Bearing Capacity and Deformation of Stored Mine
Wastes
Conventional mechanical dewatering involves the application of
different pressures to drive moisture from the capillaries of the Bearing capacity is assessed based on the shear strength of a soil.
different product coal size fractions, the pressure required being The shear strength of a saturated or desiccated tailings profile is best
dependent on the type of coal, and its pore size distribution and assessed by in situ vane shear strength testing, since tailings are
connectivity. The smaller the pores, the higher the matric (capillary) virtually impossible to sample undisturbed for laboratory testing
suction that can develop as water retreats into them, and the higher purposes.
the pressure required to remove the water. The shear strength of spoil, waste rock and coarse-grained
Generally, the dewatered moisture content of each product coal processing wastes is difficult to assess due to their coarse particle
size fraction correlates well with its particle size distribution, which size distribution, and only scalped samples can be tested in the
relates to its pore size distribution; the dewatered moisture content laboratory. Weathered spoil and waste rock, and fine-grained
increasing with decreasing particle size. Following mechanical processing wastes, are more readily tested due to their finer particle
dewatering, the coarse and fine product coal size fractions are size distribution, making scalped samples more representative. The
unsaturated, while the ultra-fine product coal size fraction remains scalped spoil, waste rock or coarse-grained processing wastes may
near-saturated to slightly unsaturated. be tested at their as-sampled moisture content, to provide an
A coarse coal centrifuge applies a pressure of 40 to 50 kPa, estimate of the unsaturated shear strength and bearing capacity, or
depending on the coal properties, the centrifuge and its operating saturated, to provide an estimate of their worst case shear strength
parameters, and the residence time of the coal within the centrifuge and bearing capacity.
(typically only 1 to 2 s, which effects the bed depth and therefore the The deformation under load of tailings, spoil, waste rock or
pressure differential across the bed). Despite the very short coarse-grained processing wastes may be assessed by
residence time, the relatively high hydraulic conductivity of the consolidometer testing. Desiccated tailings could be tested at their
coarse product coal allows it to drain rapidly. as-sampled moisture content, or the tailings could be re-constituted
A fine coal centrifuge applies a pressure of 50 to 750 kPa, and tested from a slurry.
depending on similar factors to the coarse coal centrifuge over a To assess their deformation under loading, scalped samples of
residence time of 5 to 15 s. This short residence time is insufficient spoil, waste rock or coarse-grained processing wastes could be
to allow much of the induced excess pore water pressure to tested, preferably in a large-size consolidometer, initially prepared
dissipate. However, the fine product coal “remembers” the applied in a loose state. They could be tested at their as-sampled moisture
pressure and continues to drain in a dry atmosphere. content, in a water bath to assess their collapse on wetting-up and
Conventional vacuum belt filtration of ultra-fine product coal over time, and on wetting-up to assess weathering.
applies a pressure of 60 to 80 kPa. Depending on the residence time
on the belt (typically 1 to 2 min), which is a function of its length 3.5 Performance of Mine Waste Geo-Covers
and speed, vacuum filtration drains about half of the moisture from Cover systems have evolved from the desire to limit potential
the ultra-fine product coal. Pressure filtration of ultra-fine product environmental impacts from stored mine wastes, including waste
coal applies a pressure of about 1,500 kPa. Briquetting may provide rock, coarse-grained mineral processing wastes, and fine-grained
a maximum pressure of 50,000 kPa or higher, but for not more than mineral processing wastes or tailings (Williams, 2011). The key
1 s, resulting in only about 1,500 kPa “felt” by the soil skeleton. means of limiting potential environmental impacts are to limit
Table 1 indicates the typical composition and total moisture oxidation of the stored mine wastes, and/or to limit the transport of
content of the various product coal size fractions. any oxidation products to the environment via water or air. The
most effective means of limiting oxidation of mine wastes is to store
Table 1 Typical composition and moisture of them permanently below water; however, this is generally not
product coal size fractions possible in moisture-deficit climates such as exist at many mine sites
Vacuum in Australia and elsewhere in the world. In such climates, geo-
Coarse Fine covers comprising soil and/or rock are required, applied to relatively
Filtration of
Parameter Coal Coal flat surfaces.
Ultra-Fine
Centrifuges Centrifuges In dry or seasonally dry climates, the geo-cover system should
Coal
% of total dry prevent exposure of the stored mine wastes to air-borne
50 to 70 12 to 20 12 to 20 mobilisation, and is best aimed at limiting net percolation of rainfall
mass
Total into the waste to limit the transport of any oxidation products. In
5 to 7 13 to 15 24 to 32 wet climates, the aim of the geo-cover system should be to either
moisture
87
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 43 No.1 March 2012 ISSN 0046-5828

shed rainfall runoff or, where acid neutralising materials are products, are described. These include (i) the wetting-up and drain-
available, to place a thick alkaline cover over the wastes and rely on down of initially dry surface waste rock dumps, (ii) the deposition,
the alkalinity produced to neutralise any acidity generated by the desiccation, rewetting and loading of mine tailings, (iii) the
underlying mine wastes. dewatering of product coal, (iv) the shear strength and
Geo-covers need to be understood in the context of the recharge compressibility of clay-rich coal mine spoil, and (v) the performance
of natural ground systems. The impact of tailings slurry deposition of a store and release cover placed on the top of waste rock dumps
and waste rock (or coarse-grained processing waste) dumps as they by way of rehabilitation.
wet-up due to rainfall are then superimposed on the net percolation
of a natural dry climate system. A range of geo-cover systems is 3

RAINFALL / EVAPOTRANSPIRATION
available, of variable effectiveness in limiting net percolation into
the underlying mine wastes. 2.5
Kalgoorlie, WA, average annual rainfall = 250 mm
Being located on the surface and of limited thickness, geo-
covers are in the active moisture zone, and exist mainly in an 2
Net evapotranspirative
unsaturated state. They undergo seasonal drying and wetting cycles. Net infiltrative
Moisture
Data on the recharge of natural systems in Southern Africa 1.5 excess
(Beekman et al., 1996; reproduced as Figure 12) show that in arid
1
climates, with an average annual rainfall of up to 250 mm, average
annual recharge is in the range from about 1 to 10 mm (0.4 to 4% of
0.5
250 mm). Natural systems in semi-arid climates, with an average
annual rainfall of up to 500 mm, experience an average annual 0
recharge of about 0.5 to 50 mm (0.1 to 10% of 500 mm). For 0 200 400 600 800 1000 1200 1400 1600 1800 2000
average annual rainfalls above 500 mm, there is a tapering off of RAINFALL (mm/year)
average annual recharge at a maximum of about 100 mm, implying
that an increasing proportion of rainfall reports as runoff for Figure 13 Identification of net positive
increasingly wet climates. water balance (or wet) climates

4.1 Wetting-Up of and Seepage from Waste Rock Dumps

Rainfall infiltration into a bare, loosely-dumped waste rock dump is
commonly assumed to be approximately 50% of average annual
rainfall, depending on the climate, and the physical and chemical
characteristics of the rock. However, there is a lack of field data to
confirm this. Taking Kalgoorlie in Figure 13 as a basis, the
potential impact of the wetting-up of a waste rock dump in that
climate to the point of continuum breakthrough (of half the annual
rainfall) is shown in Figure 14. Wetter climates would be expected
to result in correspondingly greater volumes of continuum
breakthrough, and this would occur more rapidly the higher the
average annual rainfall.

3
SEMI
RAINFALL / EVAPOTRANSPIRATION

ARID ARID
2.5
Kalgoorlie, WA, average annual rainfall = 250 mm

2
Net evapotranspirative

Net infiltrative
Moisture
Figure 12 Data on the recharge of natural systems in Southern 1.5 excess
Africa (after Beekman et al., 1996)
1
Potential impact of waste rock dump wetting-up
The Australian Bureau of Meteorology website
0.5
(http://www.bom.gov.au/) report data on average annual rainfall and
estimated actual average annual evapotranspiration across Australia. 0
Figure 13 is based on these data, in which the average annual 0 200 400 600 800 1000 1200 1400 1600 1800 2000
rainfall normalised by the estimated actual average annual RAINFALL (mm/year)
evapotranspiration is plotted against the average annual rainfall. For
dry climates the estimated actual average annual evapotranspiration Figure 14 Potential impact of waste rock dump
approaches the average annual rainfall, while for wet climates there wetting-up in Kalgoorlie’s climate
is a substantial excess of rainfall over actual evapotranspiration (or a
net positive water balance). In 2006, a 15 m high trial waste rock dump covering 0.7 ha was
Figure 13 allows the net positive water balance (or wet) climates constructed at Cadia Hill Gold Mine in New South Wales, Australia.
to be identified. Also marked in Figure 13 is the average annual It was instrumented with two lysimeters at the perimeter-bunded
rainfall for Kalgoorlie in Western Australia of about 250 mm, which (for the safety of plant, and also to limit runoff over the dump sides)
delineates net evapotranspirative from net infiltrative climatic surface of the dump and 24 lysimeters at the base of the dump, to
conditions. monitor rainfall infiltration through the top and base seepage
beneath the top surface and the side slopes of the dump,
4. SOME MINING APPLICATIONS OF UNSATURATED respectively. The lysimeters have been monitored for 4 years,
SOILS MECHANICS providing information of surface infiltration, storage within the
dump, and seepage from the base of the dump.
In the following sections, selected applications of unsaturated soil
mechanics to mining and mineral processing wastes, and to mineral
88
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 43 No.1 March 2012 ISSN 0046-5828

Heavy rainfall occurred during the construction of the trial

dump, resulting in a high average volumetric moisture content for May 2006 – End of construction
the placed waste rock of about 0.11 (gravimetric moisture content of
6.2%, for a measured specific gravity of 2.65), saturating about a
third of the porosity of about 0.33 (corresponding to an initial dry
density of 1.78 t/m3).
Over the 4 years of monitoring to 29 January 2010, the single
largest rainfall event occurred on 18 October 2006, with 73.8 mm Note differential settlement- November 2007
recorded in 24 hours. The annual rainfall totals recorded on Cadia’s induced ponding
trial waste rock dump have been 606 mm, 927 mm, 1,553 mm, and
1,609 mm, for 2006/2007, 2007/2008, 2008/2009 and 2009/2010,
respectively. The first two years of the monitoring period had up to Lysimeters
the average annual rainfall for the site of about 900 mm, while the
final two years of monitoring had substantially above the average
annual rainfall. Figure 16 Differential settlement and hard-panning of surface of
Surface infiltration into the flat top of the trial waste rock dump Cadia’s trial waste rock dump over time, causing internal runoff
has been monitored from 1 June 2006, following the construction of and concentrated ponding
the dump and the installation of the two lysimeters on its traffic-
compacted top surface. The lysimeters responded within 24 hours
of the first rainfall, building up to an average peak infiltration of
86% of cumulative rainfall in mid-August 2006, followed by a
gradual decline since, with little further infiltration recorded after 1
December 2006, as seen in Figure 15 (Rohde et al., 2011). The
average measured surface infiltration over the first 12 months of the
monitoring period was about 50% of cumulative rainfall, dropping
to an average 22% of cumulative rainfall over 4 years.

Figure 17 Average base seepage beneath flat top surface and angle
of repose side slopes of Cadia’s trial waste rock dump, expressed as
a % of cumulative rainfall

Maximum cumulative base seepage of about 5% of cumulative

rainfall beneath the flat top and about 10% beneath the AOR slopes
of Cadia’s trial waste rock dump occurred in early 2009, which
appears to be a consequence of heavy rainfall events in the months
prior to this.
Figure 15 Surface infiltration into Cadia’s trial waste rock dump, When base seepage has been recorded, lysimeters beneath the
expressed as a % of cumulative rainfall with time flat top of Cadia’s trial waste rock dump have flowed for about
twice as long (40 to 120 days, implying an average hydraulic
Differential settlement and hard-panning of the surface of the conductivity during drain-down of 4.3 x 10-6 to 1.4 x 10-6 m/s) as
dump (see Figure 16), and the concentration of rainfall runoff in those beneath the angle of repose slopes (20 to 60 days, implying an
ponds and at sinkholes not intersected by the surface lysimeters, are average hydraulic conductivity during drain-down of 8.7 x 10-6 to
the main factors behind the lack of recorded infiltration after 1 2.9 x 10-6 m/s), due to the greater height of waste rock involved.
December 2006. The patterns of infiltration vary markedly across Over the 4-year monitoring period, as wetting-up fingers
the dump and with variations in rainfall over time, as would penetrated Cadia’s trial waste rock dump (as shown in the
evaporation. Hence, the average recorded infiltration over 4 years laboratory; see Figure 10), the rainfall amount required to trigger
of 22% of cumulative rainfall represents a lower bound. The actual some base seepage reduced from about 30 mm initially to less than
average infiltration may be of the order of 50% of cumulative 2 mm, and the lag time (delay) before the emergence of base
rainfall. seepage reduced from about 15 days to less than 2 days (see Figure
Figure 17 shows the average cumulative base seepage 18; Williams and Rohde, 2009).
(expressed as a % of cumulative rainfall) recorded by lysimeters Table 2 presents the average initial and 4-year moisture states of
beneath the flat top and the angle of repose (AOR) slopes of Cadia’s Cadia’s trial waste rock dump. After 4 years, the degree of
trial waste rock dump. The base seepages responded broadly to saturation is starting to approach 60%, at which continuum
rainfall, increasing over time as the dump wet-up. Over the 4-year breakthrough would be expected for well-graded waste rock.
monitoring period, the average base seepage beneath the flat top of The time for a waste rock dump to reach continuum
the trial waste rock dump was about 2.7% of cumulative rainfall, breakthrough may be estimated using the continuum approach
while that beneath the AOR slopes was about 6.6%. This latter described by Williams (2006). A continuum approach is required,
value is higher, as might be expected given the lower average height since preferred flow pathways are unknown. As a result, no base
of the AOR slopes and their more open-textured surface. Both seepage is calculated until continuum breakthrough is reached.
values are very much smaller than the average surface infiltration While in reality some base seepage via preferred pathway flow does
over the same 4-year time period of 22% of cumulative rainfall or occur prior to continuum breakthrough, the amount is limited and
higher, implying that the majority of rainfall infiltration went into short-lived following significant rainfall events.
storage within the dump.
89
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 43 No.1 March 2012 ISSN 0046-5828

40 40 500

CONTIN. BREAKTHRU (mm/year)

Flat top lysimeter rainfall trigger
Angle of repose lysimeter rainfall trigger 450
35 35 Assumed rainfall infiltration of 450 mm
Flat top lysimeter delay
TRIGGER RAINFALL (mm)

Angle of repose lysimeter delay 400 (equivalent to 1.5 x 10-8 m/s)

30 30

DAYS UNTIL FLOW

350
25 25 300

20 20 250
200 January 2010
15 Trend line 15
for trigger rainfall 150
10 10
100
Trend line Possible actual
5 for delay 5 base seepage
50

0 0 0
0 100 200 300 400 500 600 700 800 900 1000 0 1 2 3 4 5 6 7 8 9 10
YEARS
CUMULATIVE RAINFALL (mm)

Figure 19 Estimated wetting-up and continuum breakthrough of

Figure 18 Trigger rainfall and delay for some base seepage
Cadia’s trial waste rock dump
following a rainfall event on Cadia’s trial waste rock dump
2000
Start for 15 m height Full for 15 m height
Table 2 Average initial and 4-year moisture states of Cadia’s trial 1800

AVERAGE RAINFALL (mm/year)

Start for 30 m height Full for 30 m height
Start for 60 m height Full for 60 m height
waste rock dump 1600
Start for 120 m height Full for 120 m height
1400
Final Beneath Typical mine life ~ 20 years, during which
Parameter Initial 1200 time dump is likely to be uncovered
Flat Top Slopes 1000
Volumetric
0.11 0.13 0.17 800
water content
600
Gravimetric
400
moisture 6.2 7.3 9.5
200
content (%)
0
Degree of 0 20 40 60 80 100 120 140
saturation 36 43 56 TIME TO CONTINUUM BREAKTHROUGH (years)
(%)
Figure 20 Estimated wetting-up and continuum breakthrough of
Over the time required to reach continuum breakthrough, even waste rock dumps of different height subjected to
initially fresh waste rock is likely to weather, hence well-graded a range of average annual rainfall totals
weathered waste rock may be assumed. An estimate of its SWCC
may be obtained from its particle size distribution, dry density and 4.2 Deposition, Desiccation, Re-Wetting and Loading of Mine
specific gravity of the soil, using Fredlund et al. (1997) and the Tailings
library of data contained within the SoilVision software Taking Kalgoorlie in Figure 13 as a basis, the dramatic impact of the
(www.soilvision.com). volume of water typically discharged with tailings slurry is shown in
From the SWCC and measured saturated hydraulic conductivity Figure 21.
of the soil, the unsaturated hydraulic conductivity function of the
3
soil may be calculated using the method of Fredlund et al. (1994).
RAINFALL / EVAPOTRANSPIRATION

Using the unsaturated hydraulic conductivity function expressed in 2.5

Kalgoorlie, WA, average annual rainfall = 250 mm
terms of volumetric water content, simulations of the wetting-up of
weathered mine waste rock by rainfall infiltration may be obtained 2
Net evapotranspirative
by iteration in a spreadsheet calculation. Spreadsheet calculations Net infiltrative
Moisture
1.5
are preferred over numerical methods, which would incur excess

convergence problems due to the high suction gradients within the 1

waste rock as it wet-up. Impact of tailings slurry deposition
For the purposes of the spreadsheet simulation, Cadia’s 15 m 0.5

high trial waste rock dump was divided into 10 equal 1.5 m sub-
0
layers. It was conservatively assumed for the purposes of the 0 200 400 600 800 1000 1200 1400 1600 1800 2000
simulation that 50% of the average annual rainfall of 900 mm RAINFALL (mm/year)
infiltrated the dump over the average 123 days/year of rainfall, with
the remainder lost to evaporation. The rainfall infiltration was Figure 21 Potential impact of tailings disposal
applied to the top sub-layer, causing it to wet-up. This raised its in Kalgoorlie’s climate
hydraulic conductivity, and when it rose sufficiently to allow
breakthrough of the top sub-layer this breakthrough was then Understanding the unsaturated flow behaviour of mine tailings is
applied to the second sub-layer, and so on through successive sub- described in Figure 22, which highlights the need to measure not
layers. Continuum breakthrough was estimated to commence after only the phreatic surface but also the matric suction profile with
about 5 years, and to become fully developed after about 7 years, as depth. Matric suctions lower than the “no-flow” hydrostatic case
shown by the solid line in Figure 19. imply net infiltration, while matric suctions higher than the no-flow
Also shown in Figure 19 is the possible average seasonal case imply net evaporation.
variation in base seepage with time, and the point reached by As tailings desiccate on exposure, a surface crust forms, beneath
January 2010 after 4 years. Applying the spreadsheet simulation to which the tailings remain slurry-like (as shown in Figure 23 for coal
a range of waste rock dump heights and average annual rainfall mine tailings). The reduced hydraulic conductivity of the desiccated
totals, Figure 20 is obtained (Williams, 2008). crust limits the depth of desiccation and maintains the underlying
tailings wet and soft.

90
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 43 No.1 March 2012 ISSN 0046-5828

-ve PORE PRESSURE +ve

Infiltrative
Evaporative (“WET”)
(“DRY”)

No-Flow


Phreatic surface within
stored tailings Desiccated tailings
Settled tailings
Re-flooded tailings

Figure 22 Flow conditions for deposited tailings

Figure 25 Settled, desiccated and re-flooded Mt Keith tailings in

column experiment

Figure 26 shows SWCC data collected directly from cemented

and structured Mt Keith tailings in the field, from re-slurried tailings
tested in a Tempe cell in the laboratory, and from the laboratory
column test, both on initial drying from a slurry and on re-drying
following flooding.

Average lab. re-drying SWCC Average lab. re-wetting SWCC

50
Field data to 0.2 m depth Field data below 0.2 m depth
45
GRAV. MOISTURE CONTENT (%)

Lab. column drying data

40
35

Figure 23 Surficial desiccated crust over fluid coal mine tailings 30

Column drying
25 and re-drying

Stolberg and Williams (2006) described the results of a 20 Laboratory Regime Field Regime

laboratory tailings column experiment conducted on fresh 15

Tempe re-wetting
hypersaline, Western Australian, hypersaline, Mt Keith nickel 10
Tempe re-drying
sulfide tailings in a Perspex column 2 m high by 300 mm in 5
diameter (see Figure 24). The column was instrumented along its 0
height with 10 sets of sensors to measure volumetric water content 0.1 1 10 100 1000 10000 100000 1000000
MATRIC SUCTION (kPa)
and matric suction. Pan evaporation was measured adjacent to the
column and seepage from the column was measured directly. The
water balance for the column allowed the actual evaporation to be Figure 26 Field and laboratory SWCC data for Mt Keith tailings
calculated.
Figure 26 shows that tailings exhibit “stiffer” behaviour, shown
by the estimated SWCCs, if they cement and development structure
in the field, which is largely destroyed on laboratory preparation for
testing. The laboratory regime corresponds to drying after initial
deposition or flooding, while the field regime corresponds to dry
conditions and those at depth unaffected by surface flooding.
Figure 27 shows the unsaturated hydraulic conductivity
Suction
sensors functions derived using the method of Fredlund et al. (1994) from
the laboratory and field SWCC data for Mt Keith tailings shown in
Figure 26. Under a hydraulic gradient of unity, these tailings under
field conditions would pass water at a rate of about 10 -10 m/s (6 to
60 mm/year), not 10-16 m/s as is implied by the laboratory-derived
SWCC data.
Moisture sensor tube Settled tailings
Chapman et al. (2008) described a field trial involving saline,
Figure 24 Instrumentation of tailings column experiment Cosmos nickel sulfide tailings in Western Australia. In order to
track the water cycle of the tailings deposition and desiccation
The tailings were placed to a height of 1.5 m at 50% solids by cycles, towers fabricated from 25 mm box section were installed in
mass, allowed to drain-down (draining to the top and the base, advance of tailings deposition, on which piezometers, and series of
accounting for about 60% of the tailings water) and desiccate, and 16 pairs of matric suction and Time Domain Reflectometry (TDR)
were then flooded and re-dried (see Figure 25). volumetric water content sensors were fixed at 300 mm centres
(Figures 28 and 29).

91
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 43 No.1 March 2012 ISSN 0046-5828

MATRIC SUCTION (kPa)

0.1 1 10 100 1000 10000 100000 1000000
1.0E-02
1.0E-04
1.0E-06
PERMEABILITY (m/s)

1.0E-08
1.0E-10 . Early deposition Further deposition
1.0E-12
1.0E-14
1.0E-16
1.0E-18 Estimated field curve

1.0E-20 Average laboratory curve

1.0E-22

Figure 27 Field and laboratory hydraulic conductivity Final deposition Desiccated tailings
functions for Mt Keith tailings
Figure 30 Cosmos tailings deposition and desiccation cycles
Deposition from 17 May to 19 July & 26 Oct to 16 Nov 2007
10
Piezometer 1@2 2@2 3@2 4@2 5@2
Matric suction sensor 9
6@2 7@2 8@2 9@2 10 @ 2

MATRIC SUCTION (kPa)

8
7
6
5
4
3
2
TDR moisture sensor
1
0
0 50 100 150 200 250
Deposition DAYS Deposition

Figure 31 Matric suction data with time during

Figure 28 Piezometer, and matric and moisture sensors
deposition and desiccation cycles

Instrumentation Figure 32 demonstrates that seepage from freshly-deposited

Tower tailings, or as a result of rainfall, will readily infiltrate through the
300 mm 16 full depth of near-saturated tailings and into the foundation (as
increments 15 evidenced by the lack of matric suction in response to 19.6 mm of
14 rainfall on 12 December 2007). However, desiccation between
13 deposition cycles drastically reduces the hydraulic conductivity of
12 the full depth of the tailings, holding up rainfall within the upper
11 0.5 m depth of tailings, from which it can readily be removed by
10
evaporation (as evidenced by the shallow wetting-up of the
9
8
desiccated tailings in response to rainfall on 28 July and 29
7 September 2007).
6 Deposition from 17 May to 19 July & 26 Oct to 16 Nov 2007

5 485
01-Jun-07 15-Jun-07 01-Jul-07
SENSOR ELEVATION (RL m)

484.9 15-Jul-07 01-Aug-07 15-Aug-07

4 01-Sep-07 15-Sep-07 01-Oct-07
484.8 15-Oct-07 01-Nov-07 15-Nov-07
3 01-Dec-07 15-Dec-07 31-Dec-07
484.7
2 484.6
19.6 mm on 12 Dec 4.8 mm on 28 July 0.2 mm on 28 Sept

1 (Foundation) 484.5
484.4

Figure 29 Sensor locations on tower and in foundation 484.3

484.2
484.1
Figure 30 shows the Cosmos tailings deposition and desiccation 484
stages over a number of months, to a total height of about 4 m. 0 10 20 30 40 50 60 70 80 90 100

Figure 31 shows the response of the matric suction sensors as MATRIC SUCTION (kPa)

they were progressively inundated by tailings, which were allowed Figure 32 Wetting-up of desiccated tailings by rainfall
to desiccate between deposition cycles.
The matric suction data shown in Figure 31 suggest that Tailings are typically allowed to desiccate and crust prior to a
desiccation following a cycle of tailings deposition is rapid (shown geo-cover being placed by way of rehabilitation. The shear strength
by the rapid rise in matric suction), wetting-up by fresh tailings of the crust is required to provide sufficient bearing capacity to
deposition is also rapid (shown by a rapid drop to zero matric safely support the equipment used to place the cover and to support
suction), and both desiccation and wetting-up extend into the the loading imposed by the thickness of cover material. The typical
foundation. The majority of the supernatant tailings water was use of large mining equipment leads to the rapid placement of an
removed by evaporation. excessive thickness of cover material over thinly-crusted tailings,
running the risk of inducing a “bow-wave” failure in the tailings.

92
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 43 No.1 March 2012 ISSN 0046-5828

This occurs if the pressure imposed by the thickness of cover The effect on shear strength of the development of matric
material extends to the softer tailings underlying the crust, and suctions on desiccation is clearly evident, as is the effect of
causes the tailings to undergo remoulding, soften, and cause bow- remoulding of the desiccated tailings. The initial effect of cover
wave failure. Figure 33 shows about 5 m of cover material placed placement was to remould and soften the tailings, causing a bow-
on thinly-crusted coal mine tailings at Ulan Mine in New South wave failure to develop. The tailings softened towards their
Wales, Australia, which caused a bow-wave failure extending about remoulded strength. Over time, drainage of the excess pore water
15 m from the toe of the cover. pressures induced by cover placement caused the tailings to
generally strengthen, particularly at depth. Towards the surface, the
low unsaturated hydraulic conductivity of the desiccated crust
resulted in only a small increase in shear strength from the initial
value on loading, as it wet-up due to drainage from the more
permeable tailings below.
The average shear strengths ranged from 2 kPa for the
underwater case, 15 kPa for the desiccated case and 7 kPa for the
remoulded case, 9 kPa immediately after cover placement (causing
bow-wave failure), and over 30 kPa after drainage.

4.3 Characterising Dewatering and Atmospheric Effects on

Product Coal
The application of pressure to a product coal size fraction induces
drainage and a transfer of the applied stress from excess pore water
pressure to an effective stress. On the removal of the applied
pressure, any transferred effective stress is “felt’ by the product coal
as matric suction, hence the two stresses are equivalent, and the
SWCC may be used to provide an estimate of the effectiveness of
dewatering under a given pressure.
Figure 35 shows the particle size distributions of typical
Figure 33 Bow-wave failure of crusted coal mine tailings due to
composite product coal and size fractions (Williams, 2009). The
rapid placement of an excessive thickness of cover material
composite product coal particle size distribution is close to that of
the dominant coarse size fraction, while the more minor fine and
Vane shear strength testing was carried out at the location
ultra-fine size fractions are largely sand-sized and silt and clay-
depicted in Figure 33 to arrive at the vane shear strength profiles
sized, respectively.
with depth shown in Figure 34 (Williams and Morris, 1987). Figure
34 shows the “underwater” shear strength profile that would result if
the tailings had always been kept below water, and is the result of CLAY SILT SAND GRAVEL COBBLES

the self-weight of the tailings, determined by its specific gravity. 100

The specific gravity of coal tailings is of the order of 1.8 to 2.2, 90

resulting in an underwater shear strength increase of about 80

70
0.8 kPa/m depth. In fact, the Ulan tailings were allowed to desiccate
% PASSING

60
periodically.
The “desiccated” and “remoulded” shear strength profiles were 50
40
measured away from the influence of the cover loading, the former Composite
30
being the peak vane shear value, and the latter the value obtained on Coarse (DMC)
20 Coarse (WOC)
remoulding the tailings by rotating the shear vane a full three Fine
revolutions. The “loaded-initially” shear strength profile was 10
Ultra-fine
0
measured through the bow-wave failure, immediately after cover 0.001 0.01 0.1 1 10 100
placement. The “loaded-drained” shear strength profile was PARTICLE SIZE (mm)
measured through the bow-wave two weeks after cover placement,
when the excess pore water pressures induced by loading had fully
dissipated. Figure 35 Particle size distributions of typical composite product
coal and size fractions (DMC = dense medium cyclone, WOC =
dense medium bath + water)
VANE SHEAR STRENGTH (kPa)
0 10 20 30 40 50
0 Figure 35 shows the measured SWCC drying and re-wetting
0.5
data and drying curves, fitted using the method of Fredlund et al.
(1994) and the SoilVision software, for typical composite product
1
coal and size fractions. The order of the SWCCs shown in Figure
1.5 36 is what would be expected, given the particle size distributions of
the materials. The coarse size fraction is most readily dewatered,
DEPTH (m)

2.5
followed by the composite, then the fine size fraction, and lastly the
ultra-fine size fraction.
3
The SWCC data shown in Figure 36 are the equilibrium values,
3.5 achieved after sufficient time is allowed at each level of matric
Underwater
4 Desiccated suction for the minimum total moisture content to be established.
Remoulded
4.5 Loaded-Initially These values will not be achieved on dewatering if the residence
5
Loaded-Drained time is limited.

Figure 34 Vane shear strength profiles for Ulan coal mine tailings

93
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 43 No.1 March 2012 ISSN 0046-5828

Figures 37 and 38 show the effect of residence time on the Figure 39 compares the SWCCs and data for filtered and
dewatering of fine and ultra-fine product coal, respectively. briquetted typical ultra-fine product coal. While vacuum filtration
would achieve at best a final total moisture content of about 35%,
the very much higher pressure applied by briquetting was found to
TOTAL MOISTURE CONTENT (%)

50
45 initially achieve a total moisture content of about 16%. The nominal
40 5,000 kPa pressure exerted by the rollers between which the
35 material is squeezed on briquetting is applied for not more than 1 s.
30 The briquette emerges from the rollers having an effective stress of
25
about 1,500 kPa (the stress that has transferred to the material
20
15
skeleton, as indicated by the measured matric suction shown in
10 Figure 39) and an excess pore water pressure of about 3,500 kPa
5 (the stress that is still carried by the pore water).
0
50
0.1 1 10 100 1000 10000
Filtered ultra-fine - Drying curve

TOTAL MOISTURE CONTENT (%)

MATRIC SUCTION (kPa) 45
Briquetted ultra-fine - Drying data
Composite - Drying data Composite - Wetting data Composite - Drying curve 40 Briquetted ultra-fine - Wetting data
Coarse - Drying data Coarse - Wetting data Coarse - Drying curve
Fine - Drying data Fine - Wetting data Fine - Drying curve 35
Ultra-fine - Drying data Ultra-fine - Wetting data Ultra-fine - Drying curve
30
25
Figure 36 SWCCs for typical composite product
20
coal and size fractions
15
10
60 5
TOTAL MOISTURE CONTENT (%)

0
50
0.01 0.1 1 10 100 1000 10000 100000 1000000
~ 43%
MATRIC SUCTION (kPa)
40 41 to 42%
Figure 39 SWCCs and data for filtered and briquetted
30 50% dewatering
typical ultrafine product coal
90% dewatering
20 ~ 22.5%
Under the action of the excess pore water pressure, the briquette
10 will continue to dry in the air, provided it is less humid than the pore
5-15 s 3.8 min 48 min space (which is likely), eventually achieving a total moisture content
0 of 2 to 5% (about 10 times lower than that achievable on vacuum
0.001 0.01 0.1 1 10 100
filtration), as the suction rises to the nominal pressure of 5,000 kPa
TIME (hours)
originally applied, inducing structure. If the briquette is re-wet, its
total moisture content will return to about 15%, since it will retain
Figure 37 Drying typical fine product coal under a suction of its compressed structure, while the vacuum filtered material could
55 kPa, simulating dewatering using a fine coal centrifuge re-wet to its original total moisture content of 42% in the absence of
an induced structure.
From Figure 37, the 5 to 15 s residence time of fine product coal A 11.5 m high composite product coal stockpile at its angle of
in a fine coal centrifuge will be relatively ineffective in dewatering repose of 38o was placed at an ex-plant total moisture content of
the material, achieving only 5 to 10% of the dewatering potential about 9.5%. After 8 days of drying under hot (maximum daytime
under the pressure applied. To achieve 50% and 90% of the temperatures of the order of 35oC), dry weather, the measured total
dewatering potential would require residence times of 3.8 min and moisture contents and inferred 5%, 7.5% and 10% contours were as
48 min, respectively. shown in Figure 40 (Williams, 2010). The high total moisture
contents towards the crest are due to the “hanging-up” of moisture-
60
retaining fines, while those at the base are due to gravity drainage.
TOTAL MOISTURE CONTENT (%)

50 The lowest total moisture contents are along the face of the
~ 42% stockpile, due to solar drying and the predominance of coarse
40 particles on the surface.
35 to 38%
50% dewatering
30 90% dewatering ~ 31%
12 5.5

11 12.1 12.1 12.3 11.2 10.5 9.2 5.0

20
10
7.0 6.5 6.8 5.3 5.6 3.2
9 5% total moisture content contour
10 6.7 6.2 5.7 6.1 4.6 7.5% total moisture content contour
8 10% total moisture content contour
1-2 min 2.3 min 15 min
HEIGHT (m)

0 7
4.7 7.7 7.7 7.5 6.9 7.6 7.2 7.8 3.5

0.001 0.01 0.1 1 10 100 6

TIME (hours) 5
4 7.7 8.2 7.7 7.7 7.8 7.8 6.3 7.5 6.0 5.5 3.7

3
Figure 38 Drying typical ultra-fine product coal under a suction of 2
85 kPa, simulating dewatering using vacuum filtration 1
6.3 12.3 12.9 8.5 11.1 11.9 10.0 12.1 9.5 7.4 10.5 12.6 8.2
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
From Figure 38, the 1 to 2 min residence time of ultra-fine
DISTANCE FROM CENTRELINE (m)
product coal on a vacuum belt will achieve 20 to 50% of the
dewatering potential under the pressure applied. To achieve 90% of
the dewatering potential would require a residence time of 15 min. Figure 40 Total moisture contents and approximate contours within
composite product coal stockpile after 8 days of drying

94
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 43 No.1 March 2012 ISSN 0046-5828

In the absence of natural rainfall, the same material was 350

subjected to simulated rainfall in a laboratory tank 2.4 m long, by 50 kPa
100 kPa
750 mm high by 300 mm wide. Two simulations were performed: 300 150 kPa
(i) rainfall at 24 mm/hour for 5 hours, immediately followed by

SHEAR STRESS (kPa)

sampling for total moisture content and, (ii) rainfall at 24 mm/hour 250

for 4 hours, followed by drain-down, then sampling. The total

200
moisture contents of the composite product coal from the field and
laboratory tests are given in Table 3, which highlights the dramatic 150
effects that climatic conditions can have on stockpiled product coal.
100
Table 3 Total moisture contents of composite product coal from
50
field stockpile and laboratory simulations
0
Final Total
0 1 2 3 4 5 6 7
Test Moisture SHEAR DISPLACEMENT (mm)
Content (%)
Field – After8 days of drying 6.8 (0.6 to 12.9)
Figure 42 Shear stress versus shear displacement plots for
Lab (i) – 5 hours rain, immediate sampling 11.8 (8.5 to 16.2) Jeebropilly weathered rock spoil placed initially loose and
Lab (ii) – 4 hours rain, after draw down 9.5 (5.4 to 13.3) tested in a water bath
4.4 Characterising Shear Strength and Compressibility of
SHEAR DISPLACEMENT (mm)
Clay-Rich Coal Mine Spoil 0 1 2 3 4 5 6 7
0
The shear strength of clay-rich coal mine spoil from Jeebropilly 100 kPa
Coal Mine in the Ipswich Coalfields of South East Queensland, NORMAL DISPLACEMENT (mm) 250 kPa
500 kPa
Australia, is substantially affected by whether the material is tested 0.2
at its relatively dry as-sampled gravimetric moisture content of
14.8% (corresponding to a matric suction of about 4,300 kPa), or in
0.4
a water bath (corresponding to near-saturation and near-zero matric
suction on wetting-up by rainfall infiltration and/or groundwater
recharge; Williams et al., 2011). 0.6

Figures 41 and 42 show the shear stress versus shear strain plots
from single-stage, 60 mm size, direct shear strength tests carried out 0.8
on weathered rock spoil specimens, scalped to pass 2.36 mm, placed
initially loose, and tested at the as-sampled moisture state and in a
1
water bath, respectively. The applied normal stress was limited to a
maximum of 500 kPa for the specimens tested at the as-sampled
moisture content, and to a maximum of 150 kPa for the specimens Figure 43 Normal displacement versus shear displacement plots for
tested in a water bath, the latter normal stress being limited by the Jeebropilly weathered rock spoil placed initially loose and tested at
large compression of the sample due to wetting-up and loading. the as-sampled gravimetric moisture content of 14.8%
Under 100 kPa normal stress, the specimen placed loose at the as- SHEAR DISPLACEMENT (mm)
sampled moisture content compressed rapidly by about 17%, while 0 1 2 3 4 5 6 7
the specimen placed loose in a water bath settled more slowly and 0
50 kPa
by 29% (almost twice the compression). Shearing was at a shear 100 kPa
NORMAL DISPLACEMENT (mm)

strain rate of 0.1 mm/min, which resulted in drained behaviour. 0.2

150 kPa

350
100 kPa
250 kPa
300 500 kPa
0.4
SHEAR STRESS (kPa)

250
0.6

200

0.8
150

100
1

50
Figure 44 Normal displacement versus shear displacement plots for
0
Jeebropilly weathered rock spoil placed initially loose
0 1 2 3 4 5 6 7
SHEAR DISPLACEMENT (mm)
and tested in a water bath

The shear stress vs. shear displacement behaviour for the two
Figure 41 Shear stress versus shear displacement plots for
cases was reasonably similar at low shear displacement (Figures 41
Jeebropilly weathered rock spoil placed initially loose and tested
and 42). However, there was a consistent step in the stress-strain
at the as-sampled gravimetric moisture content of 14.8%
behaviour of the specimens tested at the as-sampled moisture
content, at about 3 mm shear displacement (5% shear strain),
The corresponding normal displacement versus shear
corresponding to a consistent step in the normal displacement.
displacement plots during shearing are shown in Figures 43 and 44,
It is considered that the stress-strain behaviour up to about 5%
for shearing at the as-sampled moisture content and in a water bath,
shear strain was dominated by the applied normal stress,
respectively.
accompanied by relatively little further settlement. Beyond about
95
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 43 No.1 March 2012 ISSN 0046-5828

5% shear strain, the significant decrease in pore sizes associated TIME (min)
with the step in normal displacement is considered to have led to 1 10 100 1000 10000
increased matric suction and hence effective stress, causing an 0
20 kPa
increase in the measured shear stress required to further shear the 0.5 40 kPa
80 kPa
specimens. This effect was not observed for the specimens in a 1 160 kPa
water bath since they were essentially water-saturated, hence matric 1.5
320 kPa

% SETTLEMENT
480 kPa
suction would have been close to zero. 960 kPa
2
Figure 45 compares the direct shear strength envelopes derived
from the stress-strain plots in Figures 41 and 42. The plateaux in the 2.5

plots for the specimens tested at the as-sampled moisture content 3

and the peaks for the specimens tested in a water bath gave similar 3.5
drained shear strength parameters, with an effective stress cohesion 4
intercept of about 13 kPa and an effective stress friction angle of
4.5
about 17.9o. The final points of the plots for the specimens tested at
5
the as-sampled moisture content gave an effective stress cohesion
intercept of about 33 kPa and an effective stress friction angle of
about 27.3o. Figure 47 Settlement versus time plots for Jeebropilly weathered
rock spoil placed initially loose and tested in a water bath
350
Dry @ 100 kPa-Plateau
Dry @ 250 kPa-Plateau
Figure 48 shows the consolidation plots from the end-points of
300 Dry @ 500 kPa-Plateau the plots shown in Figures 46 and 47. Also indicted in Figure 48 are
Dry @ 100 kPa-Final the increase in dry density from 1.020 to 1.121 t/m3 (from 61 to 67%
SHEAR STRESS (kPa)

Dry @ 250 kPa-Final

250
Dry @ 500 kPa-Final of Maximum Dry Density, MDD, of 1.68 t/m3) with increasing
Wet @ 50 kPa-Peak c' ~ 33 kPa, ' ~ 27.3o applied stress when tested at the as-sampled moisture content,
200 Wet @ 100 kPa-Peak
Wet @ 150 kPa-Peak
increasing to 1.204 t/m3 (72% of MDD) on flooding in a water bath,
150 and increasing further to 1.469 t/m3 (87% of MDD) on loading in a
c' ~ 13 kPa, ' ~ 17.9o
water bath.
100

1.6 Dry density ~ 1.020 t/m3 ~ 61% of MDD

50
1.5 Cc = 0.115

0 1.4
0 100 200 300 400 500 600 700
NORMAL STRESS (kPa) 1.3
1.121 t/m3 ~ 67% of MDD
VOID RATIO

1.2 1.204 t/m3 ~ 72% of MDD

Figure 45 Direct shear drained strengths of Jeebropilly weathered 1.1

rock spoil tested dry (as-sampled moisture content) 1 Cc = 0.167

and wet (in a water bath) 0.9

0.8
The compressibility of the same Jeebropilly weathered rock As-sampled moisture state
0.7 1.469 t/m3 ~ 87% of MDD
spoil is also substantially affected by whether the material is tested In a water bath

dry or wet, as shown in Figures 46 and 47. These data were 0.6
10 100 1000
obtained by 75 mm diameter oedometer testing of -2.36 mm-
APPLIED STRESS (kPa)
scalped, initially loose-placed specimens, at the as-sampled moisture
content and in a water bath, respectively.
TIME (min)
Figure 48 Consolidation plots for Jeebropilly weathered rock spoil
1 10 100 1000 10000 placed initially loose and tested at the as-sampled
0 moisture content or in a water bath
20 kPa
0.5 40 kPa

1
80 kPa
The plots obtained for the “dry” and “wet” cases are both
160 kPa

1.5
320 kPa essentially linear, and as the material is loosely-placed represent
% SETTLEMENT

480 kPa
960 kPa
normally consolidated conditions. The values of Compression
2
Index Cc calculated for testing at the as-sampled moisture content
2.5
and in a water bath are shown in Figure 48 as 0.115 and 0.167,
3 respectively, confirming that the material tested in a water bath is
3.5 almost twice as compressible as the material tested at the as-sampled
4 moisture content.
4.5 To assess the relative magnitude of weathering-induced
5
settlement of the same Jeebropilly weathered rock spoil, loosely-
placed, -19 mm scalped material, placed at the as-sampled moisture
content, was exposed to the weather for a period of 35 days, during
Figure 46 Settlement versus time plots for Jeebropilly weathered which 112.6 mm of rainfall fell on 20 of the 35 days (ranging from
rock spoil placed initially loose and tested at the as-sampled 0.2 to 19.4 mm/day, with an average of 5.6 mm/rain day and
gravimetric moisture content of 14.8% 3.2 mm/day overall).
Figure 49 shows that the median particle size D50 generally
Figures 46 and 47 show that testing at the as-sampled moisture decreases with cumulative rainfall, although there is some reversal,
content resulted in about half the relative settlement compared with considered to be due to re-agglomeration on wetting and drying
testing in a water bath, the difference considered to be due to cycles. Figure 50 shows the substantial % settlement of up to 25%
flooding-induced “collapse” settlement and weathering of the that the spoil undergoes on weathering, with some reversal apparent
material. on re-agglomeration.
96
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 43 No.1 March 2012 ISSN 0046-5828

120 12

Cumulative rainfall

MEDIAN PARTICLE SIZE D50 (mm)

100 D50 10
CUMULATIVE RAINFALL (mm)

Log. (D50)

80 8

60 6
Store/release
40 4 cover

20 2

0 0
0 5 10 15 20 25 30 35
TIME (days)

Figure 49 Variation of median particle size D50 with cumulative

rainfall and time for Jeebropilly weathered rock spoil placed initially Figure 51 Schematic of a typical store and
loose and exposed to weathering release cover over mine wastes

TIME (days) The first application of the store and release cover system was
0 5 10 15 20 25 30 35 on the tops of mineralised waste rock dumps at Kidston Gold Mine
0 in north Queensland, Australia (Williams et al., 1997). The store
Data
Poly. (Data ) and release cover was instrumented with matric suction and
5 volumetric moisture content sensors through its thickness, and a
AVERAGE % SETTLEMENT

weather station was installed on the cover. Data collected from the
10
cover are shown in Figure 52 (Williams et al., 2006), which
demonstrates the seasonal wetting-up and drying of the cover,
15
returning to similar wet-up and dry states each wet and dry season.
The field SWCC data obtained from the Kidston instrumented
20
store and release cover are shown in Figure 53, which shows the
25
initial seasonal fingering of the data, with “equilibrium” curves
developing over time at each elevation in the cover.
30
The response of lysimeters beneath the Kidston store and release
cover has shown that net percolation has been limited to about 1%
of annual rainfall. The limited store and release cover performance
Figure 50 Settlement versus time for Jeebropilly weathered rock data available worldwide (shown in Figure 54; Williams, 2011)
spoil placed initially loose and exposed to weathering suggest that, provided appropriate materials and construction
methods are employed, they can be effective in limiting net
4.5 Characterising Performance of a Store and Release Cover percolation to between 1% and 2% of cumulative rainfall, for at
The store and release cover system was developed specifically for least 10 years.
mine sites located in dry or seasonally dry climates (Williams et al.,
1997), with the intention of developing a cover system more robust
800
than the more conventional rainfall-shedding cover; a cover system
RAIN (mm)

600
that does not rely on rainfall-shedding and the consequent risk of 400
failure of the cover. 200
0
Store and release covers comprise a compacted sealing layer S ~ 1.0, w ~ 45%
0.55
overlain by a gently-hummocked, loosely-placed rocky soil mulch Compacted Clayey layer
VOLUMETRIC WATER CONTENT

0.5
layer, which is vegetated with a mix of shrubs, trees and grasses
0.45
appropriate to the climatic setting (see Figure 51). The intention of 0.4 Wet season av
Rocky soil mulch

the compacted sealing layer is to “hold-up” and store wet season 0.35
S ~ 0.7, w ~ 20%

rainfall infiltration in the loose rocky soil mulch growth medium, 0.3
which is sized to store significant wet season rainfall infiltration, 0.25
while not being so thick that the roots of the vegetation are unable to 0.2
0.15
access it. A typical thickness of loose rocky soil mulch is 1.5 to
0.1
2 m. The stored infiltration is then released during the dry season Dry season min Sensor Depths (m)
0.05 S ~ 0.2, w ~ 5%
through evapotranspiration. The rocky soil mulch is placed by 0
0.35 0.75 1.37 1.75

paddock dumping and the surface is smoothed using a low bearing

Ju 6

Ju 7

Ju 8

Ju 9

Ju 0

Ju 1

Ju 2

Ju 3
96

D 7

D 8

D 9

D 0

D 1

D 2

D 3

04
-9

-9

-9

-9

-0

-0

-0

-0
9

pressure dozer to smear the preferred flow pathways between truck-

n-

n-

n-

n-

n-

n-

n-

n-

n-
ec

ec

ec

ec

ec

ec

ec

ec
Ju
D

dumps, leaving a gently-hummocked surface with catchments DATE

sufficiently small that erosion will be negligible. Typically, the
smoothed surface is first planted with native shrubs and trees, Figure 52 Rainfall and volumetric water content data collected from
excluding grasses since these tend to out-compete shrubs and trees. Kidston’s store and release cover
Grasses are then planted the following wet season. Aerial seeding is
preferred since it allows planting immediately following rainfall.
The resulting store and release cover system is relatively straight-
forward to construct and is robust.

97
 Geotechnical Engineering Journal of the SEAGS & AGSSEA Vol. 43 No.1 March 2012 ISSN 0046-5828

0.7
7. REFERENCES
0.35 m depth (field data)
Beekman, H. E., Gieske, A. and Selaolo, E. T. (1996) “GRES:
VOLUMETRIC WATER CONTENT

0.35 m depth (fitted curve)

0.6
1.00 m depth (field data) Groundwater recharge studies in Botswana 1987-1996”.
1.00 m depth (fitted curve)
0.5 Wet sealing layer
1.37 m depth (field data)
Botswana Journal of Earth Sciences, III, pp1-17.
1.37 m depth (fitted curve) Chapman, P. J., Williams, D. J., Rohde, T. K. and Ennor, S. J.
0.4 1.77 m depth (field data) (2008) “Understanding the water balance of potentially acid
1.77 m depth (fitted curve)
0.3 forming tailings deposited in a dry climate”. Proceedings of
Desiccated surface 3rd Int. Seminar on Mine Closure, 14-17 October 2008,
0.2 Johannesburg, South Africa, pp400-509.
0.1
Fredlund, M. D., Fredlund, D. G. and Wilson, G. W. (1997)
“Prediction of the soil water characteristic curve from grain
0.0 size distribution and volume mass properties”. Proceedings
1 10 100 1000 10000 of Third Brazilian Symp. on Unsaturated Soils, Rio de
MATRIC SUCTION (kPa)
Janeiro, Brazil, 22-25 April 1997, 12 p.
Fredlund, D. G., Xing, A. and Huang, S. (1994) “Predicting the
Figure 53 SWCC data collected from Kidston’s store permeability function for unsaturated soils using the soil
and release cover water characteristic curve”. Canadian Geotechnical Journal,
31, pp533-546.
100
Rohde, T. K., Williams, D. J. and Burton, J. (2011) “Waste rock
90 Cadia dump rainfall infiltration and base seepage”. Proceedings of
Kidston 7th Australian Workshop on Acid and Metalliferous
80 Mt Whaleback 2 m ROM-av
Drainage, Darwin, Australia, 21-24 June 2011. 17 p.
(% cumulative rainfall)
NET PERCOLATION

Mt Whaleback 2 m ROM-peak
70 Mt Whaleback 4 m ROM Stolberg, D. J. and Williams, D. J. (2006) “Large-scale column
Goldstrike
60 testing of hypersaline tailings”. Proceedings of 5th Int.
50 Congress on Environmental Geotechnics, Cardiff, Wales, 26-
40 30 June 2006, II, pp976-983.
30 Williams, D. J. (2006) “Mine closure as a driver for waste rock
20 dump construction”. Proceedings of 1st Int. Seminar on Mine
10
Closure, Perth, Australia, 13-15 September 2006, pp697-706.
Williams, D. J. (2008) “The influence of climate on seepage from
0
0 2 4 6 8 10 12 14 16 18 20
mine waste storages during deposition and post-closure”.
AGE (years)
Proceedings of 3rd Int. Seminar on Mine Closure, 14-17
October 2008, Johannesburg, South Africa, pp461-473.
Williams, D. J. (2009) “Use of SWCCs to describe the dewatering
Figure 54 Performance data on store and release covers on of product coal”. Proceedings of 4th Asia Pacific Conf. on
the tops of waste rock dumps worldwide Unsaturated Soils, Newcastle, Australia, 23-25 November
2009, pp233-238.
5. CONCLUSIONS Williams, D. J. (2010) “Atmospheric drying and laboratory wetting
There are many applications of unsaturated soil in the mining field. of stockpiled product coal”. Proceedings of 5th Int. Conf. on
These include the wetting-up and drain-down of initially dry surface Unsaturated Soils, Barcelona, Spain, 6-8 September 2010, 1,
waste rock dumps; the irrigation and drain-down of heap leach pp513-518.
materials; the drain-down, desiccation and rewetting of mine Williams, D. J. (2011) “Keynote address: appropriate geo-cover
tailings; the dewatering of mineral products such as coal; the shear systems for different climates”. Proceedings of 7th
strength and compressibility of stored mine wastes; and the Australian Workshop on Acid and Metalliferous Drainage,
performance of geo-covers placed on mine wastes by way of Darwin, Australia, 21-24 June 2011, 17 p.
rehabilitation. This paper highlights the key unsaturated soil Williams, D. J. and Morris, P. H. (1987) “Bearing capacity and
mechanics parameters involved, overviews the nature of mining and deformation characteristics of ponded fine-grained coal mine
mineral processing wastes, and some mineral products, and tailings”. Proceedings of Nat. Conf. on Mining and
discusses a number of the issues involved. Some applications of Environment - A Professional Approach, Brisbane, Australia,
unsaturated soil mechanics addressing the strength, compressibility July 1987, pp139-144.
and permeability of mining and mineral processing wastes, and Williams, D. J. and Rohde, T. K. (2009) “Reliability of using
mineral products, are presented, together with data to highlight laboratory-determined soil water characteristic data for mine
them. waste cover design”. Proceedings of Mine Closure 2009,
Perth, Australia, 9-11 September 2009, pp493-504.
6. ACKNOWLEDGEMENTS Williams, D. J., Stolberg, D. J. and Currey, N. A. (2006) “Long-
term performance of Kidston’s ‘store/release’ cover system
This paper is based on many years of research on the application of over potentially acid forming waste rock dumps”.
unsaturated soil mechanics principles to mine waste management Proceedings of 7th Int. Conf. on Acid Rock Drainage, St
and mined landform design, which has involved numerous past Louis, Missouri, USA, 26-30 March 2006, pp2385-2396.
research students and research funding primarily from the Australian Williams, D. J., Wilson, G. W. and Currey, N. A. (1997) “A cover
Research Council Linkage Program, the Australian Coal Association system for a potentially acid forming waste rock dump in a
Research Program, and from industry. dry climate”. Proceedings of 4th Int. Conf. on Tailings and
Mine Waste ’97, Fort Collins, Colorado, 13-17 January 1997,
pp231-235.
Williams, D. J., Kho, A. and Daley, A. (2011) “Settlement and
strength of clay-rich coal mine spoil”. Proceedings of
Tailings and Mine Waste 2011, Vancouver, Canada, 6-9
November 2011, 12 p.
98