predicting the performance of an elevated water table for preventing

OttawaGeo2007/OttawaGéo2007
PREDICTING THE PERFORMANCE OF AN
ELEVATED WATER TABLE FOR PREVENTING ACID
MINE DRAINAGE
1
1
1,*
2,*
1
M. Ouangrawa , J. Molson , M. Aubertin , B. Bussière , & G.J. Zagury
1
Ecole Polytechnique Montreal, Dept. of Civil, Geological & Mining Engineering
2
Université du Québec en Abitibi-Témiscamingue
*
Industrial NSERC Polytechnique/UQAT Chair in Environment and Mine Waste
Management
ABSTRACT
The elevated water table (EWT) technique can be an advantageous method for the management and closure of reactive
mine tailings impoundments. The principle consists of maintaining by submersion or capillarity a high degree of
saturation within the tailings in order to inhibit sulphidic mineral oxidation. Its potentially high efficiency is due to the low
rate of oxygen diffusion under nearly saturated conditions. The effectiveness of an elevated water table in preventing
acid mine drainage (AMD) can be quantified using numerical simulations of water flow, oxygen diffusion, sulphide
oxidation and geochemical transport. In this study, the reactive transport model MIN3P (Mayer, 1999) is applied to
simulate several laboratory columns with different water table elevations. Each column contained reactive tailings
overlain by a sand layer and was subjected to transient recharge and drainage (fifteen 30-70 day cycles for 502 days).
A free-draining control column without a sand layer was also used. The numerical results were generally consistent with
the observed experimental data, which showed more neutral pH and lower SO4 and Fe concentrations as the water
table elevation was increased. The results highlight the conditions under which EWTs can significantly reduce AMD
generation.
RÉSUMÉ
La technique de la nappe phréatique surélevée (NPS) peut constituer une méthode avantageuse pour la gestion et la
fermeture de parcs à résidus miniers réactifs. Le principe consiste à maintenir les résidus à un degré de saturation
élevé par submersion ou par capillarité afin d'empêcher l'oxydation des minéraux sulfureux. La technique repose sur le
faible taux de diffusion de l'oxygène en milieu quasi saturé. Cette efficacité à empêcher le drainage minier acide (DMA)
peut être estimée à partir de simulations numériques de l’écoulement de l'eau, de la diffusion de l'oxygène, de
l’oxydation des sulfures et du transport géochimique. Dans cette étude, le modèle de transport réactif MIN3P (Mayer,
1999) est utilisé pour simuler des essais en colonnes avec différentes positions de la nappe. Chaque colonne comporte
des résidus sulfureux surmontés par une couche de sable. On y applique une recharge avec drainage (15 cycles de 3070 jours pendant 502 jours). Une colonne témoin avec un drainage contrôlé et sans couche de sable a été également
utilisée. Les résultats numériques sont généralement conformes aux données expérimentales observées qui montrent
que plus le niveau de la nappe est élevé dans les résidus, plus le pH est élevé et les concentrations en sulfates et en
fer sont faibles. Ces résultats montrent les conditions requises pour que la technique de la NPS puisse réduire la
production de DMA de manière significative.
1
INTRODUCTION
The Canadian mining industry is of critical importance to
the country’s economy. However, ore processing
produces significant volumes of potentially reactive waste
tailings, which must be carefully managed to prevent
environmental impacts including acid mine drainage
(AMD). Mine tailings are usually deposited in surface
impoundments surrounded by dykes and dams. In such
environments, and when sulphidic minerals are present in
the tailings, water covers (or underwater disposal) are
often used to prevent AMD. The use of water covers for
controlling AMD has been well documented over the
years (e.g. David and Nicholson, 1995; Li et al., 1997;
Amyot et Vezina, 1997; Catalan et al., 2000; Adu-Wusu
et al., 2001; Vigneault et al., 2001; Peacey and Yanful,
2003; Mian and Yanful, 2003). Although effective, water
covers can be expensive to maintain and can lead to
geotechnical instability (Aubertin et al. 1997).
1470
Over the last decade, the elevated water table (EWT)
technique has received increasing interest as a possible
option for management and closure of reactive tailings
impoundments (MEND, 1996). The principle of the
technique consists of maintaining a high degree of
saturation within the tailings by submersion and/or
capillary rise in order to inhibit sulphide oxidation. Its
efficiency is based on the low rate of oxygen diffusion in
saturated or near-saturated conditions. Oxygen diffusion
rates through highly saturated porous media, for example,
can be several orders of magnitude less than under
relatively dry conditions (Nicholson and Tibble 1995;
Aubertin et al., 1995; 2000; Mbonimpa et al. 2003; Aachib
et al., 2004). The effectiveness of the EWT technique
depends on many factors including the capillary retention
properties of the waste material, climatic conditions, water
balance of the impoundment, as well as the tailings
mineralogy and buffering capacity (Ouangrawa et al.
OttawaGeo2007/OttawaGéo2007
2005; 2006; Dagenais et al. 2006). For optimal efficiency,
the depth of the water table must be less than the air
entry value (AEV) of the tailings (Orava et al., 1997;
Aubertin et al., 1999; Ouangrawa et al. 2005; 2006;
Dagenais et al. 2006). Compared to water covers, an
elevated water table can yield about the same low
sulphide oxidation rates, can cost significantly less, and
can improve the geotechnical stability of the engineered
structure. The main drawback of an EWT is that data are
scarce with regard to its effectiveness. Also, the optimal
design conditions have not yet been well defined. In this
regard, Dagenais (2005) used laboratory column
experiments for evaluating the performance of the EWT
technique (see also Dagenais et al. 2006). This past
research has lead to an improved understanding of how
the hydrogeological conditions affect the efficiency of the
EWT technique.
This study focuses on hydrogeochemical analyses and
numerical modelling to evaluate the effectiveness of an
elevated water table for controlling acid mine drainage
from sulphidic mine tailings. The laboratory experiments
used columns of tailings from the Louvicourt mine in
Quebec, Canada. The columns had different water table
elevations and were subjected to transient recharge.
Water flow, oxygen consumption, and chemistry of the
drainage water were monitored to establish the
hydrological and geochemical conditions which control
system behaviour. The observed behaviour is compared
with numerical reactive transport simulations. The latter
are used to evaluate how an EWT may be effective in
preventing the production of AMD. More details on this
study can be found in Ouangrawa (2007).
2
MATERIALS AND LABORATORY SET-UP
The laboratory study included a total of 15 short
cylindrical experiment columns (0.4 m long) with different
tailings material and/or negative (suction) pressures at
the base. In this paper, the authors will consider only four
of these test columns (columns 1, 4 & 12), which were
prepared using tailings sampled from the Louvicourt mine
site, located near Val-d’Or, Quebec, Canada. Larger
columns (1.5 m to 2 m long) were also tested but will not
be considered here.
The general set-up for the short column tests is
represented in Figure 1. Each column included 15 cm of
consolidated tailings and 20 cm of sand (acting as an
evaporative barrier). At the top, the columns were open to
the atmosphere. A removable TDR probe was used to
measure volumetric water content in the sand. A
removable sensor fixed to a cap was also used to
measure oxygen consumption in the columns which were
temporarily closed after three weeks of drainage;
however, the oxygen-consumption test results are not
presented here (see Ouangrawa 2007 for details). At the
base, the columns were equipped with a ceramic porous
plate. The water table position was adjusted by applying
suction through a port connected to a U-shape plastic
tube. The tube was also used to collect leachate for
chemical analyses.
1471
Figure 1. Typical experimental column and
instrumentation
The initially unoxidized Louvicourt mine tailings are
composed primarily of silt and fine sand size particles
with a percentage passing 80 µm of about 90% (and 80%
passing 40 µm). Characteristic grain diameters,
corresponding to 10% (D10), 50% (D50), and 60% (D60)
passing on the cumulative grain size distribution curve,
are 2 µm, 17 µm and 24 µm, respectively.
The main observed silicate minerals are quartz (19 wt.
%), chlorite (11 wt. %), feldspars (1.5 wt. %) and
muscovite (11 wt. %). Pyrite (38 wt. %) is the main
sulphide mineral with small amounts of chalcopyrite and
sphalerite. The main carbonates are dolomite (3 wt. %)
and siderite (7 wt. %). The saturated hydraulic
-6
-5
conductivity of the tailings is in the range 10 - 10 cm/s
at a void ratio of 0.5 - 0.7, and the Air Entry Value (AEV)
is approximately 300-350 cm H2O.
The chemical composition of the tailings was determined
by Inductively Coupled Plasma analysis (ICP-AES) after
acid-Bromine digestion. Dilute HCl was used to extract
sulfates and the resultant solution was analyzed by ICPAES. The silica content was determined by ICP-AES
analysis following a Na2O2/NaOH fusion. The results
show that the Louvicourt mine tailings are rich in Fe (27.3
wt%) and S (19.5 wt%) due to the presence of pyrite.
Total sulphur and total inorganic carbon were obtained
with the LECO CS-400 Carbon/Sulfur Series apparatus.
The results were consistent with the ICP-AES analyses.
The Louvicourt mine tailings were selected partly
because of their high acid generating potential. Acid-base
accounting (ABA) tests (Lawrence et Wang, 1997)
showed that the tailings net neutralization potential (NNP)
is approximately -558 kg CaCO3/tonne and the acid
generation potential (AP) is 599 kg CaCO3/tonne.
In columns 1 and 4, suctions of 209 cm and 130 cm H2O,
respectively, were applied to the base. These different
suctions represent different water table depths below the
tailings surface. Column 12, used for control, had neither
OttawaGeo2007/OttawaGéo2007
a ceramic plate nor a sand cover; the tailings were
underlain by a geotextile. The equivalent suction in
column 12 was zero from 0-322 days (cycle 1-8) and then
was increased to about 10 m (according to the WRC)
from 322-502 days (cycle 9-15) by opening the base to
free drainage.
Once a month, each column was wetted with
3
approximately 570 cm of demineralised water (pH~6)
added at the upper surface. The columns were then
allowed to drain. A total of 15 drainage cycles were
completed, each lasting approximately 30 days (cycle 1
lasted 70 days). The pH and redox potential of the
leachate were measured after each cycle, while the
different concentrations (SO4, Mg, Mn, Ca, Zn, Cu, Fe…)
in the leachate were determined by atomic absorption
spectroscopy.
saturation. The top boundary was a Type-2 (Neumann)
flux condition with cyclic recharge events. At the
-7
beginning of each cycle, a recharge of 6.4×10 m/s (= 2.3
mm/hr) was applied for approximately 3 hours, followed
by a zero-flux condition for the remainder of the cycle.
The chemistry of the recharge water for each simulated
cycle was based on measured concentrations of the
demineralized water used in the experiments
(Ouangrawa, 2007).
Table 1. Material characteristics used in the MIN3P
simulations.
nv
AEV
Material ksat
θr
αv
-1
[m]
[m/s]
[m ] [-]
1
Lou
Sand
3
2
CP
NUMERICAL MODELLING
1
3.1
MIN3P code
2
The reactive transport simulations were conducted using
the MIN3P code (Mayer, 1999; Mayer et al. 2002). MIN3P
is a finite volume based numerical model for simulating
3D variably-saturated flow, advective-dispersive transport
in the water phase, diffusive gas transport, sulphide
mineral oxidation, and multi-component kinetic or
equilibrium controlled geochemical reactions including
mineral buffering. The model is based on a system of
nonlinear equations for coupled flow and mass transfer in
soil, and has been used extensively in previous studies of
reactive mine wastes, including those by Bain et al.
(2001), Mayer et al. (2002), Jurjovec et al. (2004), and
Molson et al. (2004, 2007).
3.2
Modelling methodology
Physical and hydro-geochemical material characteristics
determined from the laboratory (Table 1) were used as
input data to the MIN3P model to simulate the
experiments. The numerical model considers three layers
(from the bottom to the top): the ceramic plate (0.6 cm),
the tailings (15 cm), and the sand (20 cm) as shown in
Figure 1. The van Genuchten (1980) model parameters
(αv and nv) were determined using the RETC code
-7
according to van Genuchten et al. (1991). A ksat of 10
m/s and an AEV of 3.50 m of water were used in the
simulations. The AEV and the residual volumetric water
content (θr) were also measured and determined with the
modified Kovacs (MK) model (Aubertin et al. 2003). At the
base of each column, suctions were applied which
corresponded to those used in the laboratory experiment.
The hydraulic characteristics (AEV and ksat) used for the
ceramic plate are given by the manufacturer (Soilmoisture
Equipment Corp, 2002).
Each column was simulated as a vertical 1-D system,
with columns 1, 4 and 6 resolved with 200 elements and
column 12 with 90 elements (each 1.83 mm high). The
bottom flow boundary condition was a Type-1 (Dirichlet)
with a fixed negative water pressure to control the tailings
1472
1.0×10
-7
0.10
0.10
2.60
3.50
1.0×10
-3
0.05
5.32
25.0
0.14
3.1×10
-7
0.10
0.05
3.00
5.00
Louvicourt tailings
Ceramic porous plate
Each column was simulated for a 502 day period, with
-4
time steps varying from a minimum of 2×10 s to a
maximum of 0.8 days.
The initial mineralogy and aqueous phase chemistry of
the tailings were based on observed laboratory data (see
Section 2). The secondary minerals which were allowed
to precipitate in the model were: siderite, gibbsite,
gypsum, ferrihydrite, and jarosite. Sixteen primary
+2
+1
-1
+3
aqueous components (Ca , K , Cl , H4SiO4, Al , CO3
2
+1
+2
-2
+3
-1
+2
2+
2+
, H , O2(aq), Fe , SO4 , Fe , HS , Mg , Zn , Cu )
and 38 secondary aqueous species were included in
each simulation. A dispersivity of 0.5 mm was assumed
for all simulations and the materials in the columns were
considered initially fully saturated (Sw = 100%). Execution
times were typically on the order of 3 hours to 2-3 days
on a Pentium IV, 2GHz machine.
4
RESULTS AND DISCUSSION
4.1
Hydrogeochemical
simulated)
evolution
(observed
and
The main results obtained from the laboratory column
tests and from the numerical simulations are analyzed
and compared in relation to the performance of the EWT
as a method to limit AMD production. Figures 2, 3 and 4
present the simulated vertical profiles of the degree of
2+
saturation Sw, dissolved oxygen, pH, Eh, sulphate, Ca
and Fe at selected times for columns 12, 1 and 4,
respectively.
OttawaGeo2007/OttawaGéo2007
Column 12 = control
saturation
O2(aq)
Eh
pH
Ca2+
SO4
0.4
Column 1
suction = 2.094 m
saturation O2(aq)
Total iron
0.4
pH
0
0d
502
502 d
1h
0
0
0 0.5 1 0 5
mg/L
Sw
0 day
322
0 71 days
0 day
jr
1h
502
71
322
71
Elevation (m)
0 day
0.2
0.1
71
0.3
502
0 day
71
502
322
322
502
5 10 0
1000 0100000 0 500
mg/L
pH
mV
mg/L
0.2
1day
1 day
2d
502
322
0
Figure 2. Simulated vertical profiles for control column 12
(suction = 0 to 1000 cm) showing degree of saturation,
oxygen, pH, Eh, sulphate, calcium and iron
concentrations at selected times (time in days unless
otherwise noted).
In the controlled drainage case (column 12), the
simulation results (Figure 2) show that when the suction
applied at the base is changed from zero to about 10 m
H2O, the degree of saturation of the tailings decreases
from 100% to approximately 50%. The desaturation
allows oxygen to penetrate and reach the bottom of the
column. As the tailings desaturate, the sulphide minerals
begin to oxidize which releases acidity. This is reflected
by a decrease of pH (from 7 to 3 at the end of the
experiment) and an increase of Eh (from 200 mV to 1000
mV). This simulated Eh value is higher than that
measured which is about 600 mV. At the same time, the
simulated sulphate concentrations increased from
approximately 3000 mg/L to nearly 100,000 mg/L, iron
from 0.1 mg/L to about 50,000 mg/L, and calcium from
less than 100 mg/L to 500 mg/L.
At the beginning of each cycle for columns 1 and 4, water
drains rapidly from the sand cover, but the tailings remain
close to full saturation (degree of saturation Sw ~0.9 for
column 1 and 0.95 for column 4). As the sand layer
desaturates, oxygen diffuses rapidly into the column, but
diffusion into the underlying tailings is limited because of
its higher saturation. The limited oxygen availability allows
the tailings to partially oxidize at the sand/tailings
interface. The pH drops locally to about 2-3 by the end of
the experiment (502 days). A limited oxygen supply
(because of the higher water saturation) combined with
pH buffering (by dolomite and calcite) limits the oxidation
and prevents this low pH front from advancing further
down the column. The pH at the base therefore remains
around 6-7. The sulphate and iron concentration profiles
can be correlated to the pH drop at the sand/tailings
interface (see Figures 3 and 4 below).
The extent of localized oxidation is somewhat less in
column 4 because the applied suction at the base of this
column was less and the tailings therefore remained
closer to full saturation (Sw~0.95). As a result, the
sulphate and iron concentrations are also significantly
lower.
1473
71
322
322
71
322
2d
322
502
322
0 50000
mg/L
0d
71 d
5h
0.1
502
0 0.5 10 5 0
mg/L
Sw
502
502
502
71
5 10 0
1000
pH
mV
0 10000
mg/L
0 1000 0 2000
mg/L
mg/L
Figure 3. Simulated vertical profiles for column 1 (suction
of 209 cm) showing degree of saturation, oxygen, pH, Eh,
sulphate, calcium and iron concentrations at selected
times (time in days unless otherwise noted).
Column 4
suction =1.30 m
saturation O2(aq)
0.4
pH
0
0 day
71
0.3
1day
0.2
322
Ca2+
0 day
Total iron
0 day
71
502
71
322
322
71
322
2d
0.1
0
0 day
5h
1 day
2d
SO4
Eh
5h
Elevation (m)
Elevation (m)
0.3
0d
0d
0 day
5h
Total iron
2+
Ca
SO4
Eh
502
0 0.5 10 5 0
mg/L
Sw
502
5 10 0
1000
pH
mV
322
502
502
71
0 10000
mg/L
0 1000 0 2000
mg/L
mg/L
Figure 4. Simulated vertical profiles for column 4 (suction
of 130 cm) showing degree of saturation, oxygen, pH, Eh,
sulphate, calcium and iron concentrations, at selected
times (time in days unless otherwise noted). There is less
sulphide oxidation in this case compared to column 1
(Fig. 3) because the lower suction at the base of column
4 maintains a higher level of water saturation.
The observed effluent data for the 15 drainage cycles
(about 500 days) are compared to the simulated data in
Figure 5. In column 12, the geochemical evolution of the
tailings can be clearly related to the change in free
drainage at 322 days. In general, similar trends are
observed for the laboratory and simulated data. The pH
remained between 7 and 8 for the first 322 days, and then
gradually decreased to 4.5 by 500 days. The highest
sulphate, iron, zinc and copper concentrations observed
in control column 12 can be related to oxidation of pyrite,
sphalerite and chalcopyrite.
The simulated pH in column 1 is somewhat less than the
measured pH as shown in Figure 5. The simulated
sulphate concentrations in column 1 remained between
2000-6000 mg/L throughout the study period, whereas
the observed data are between 2000-3000 mg/L. The
same observation can be made for iron, zinc and copper
for which the simulated concentrations are higher than
those observed (see Figure 5).
OttawaGeo2007/OttawaGéo2007
In column 4, the observed sulphate concentration
gradually decreased from about 1500 mg/L to 400 mg/L
after 500 days. This trend is not observed in the
simulated data, which remained higher than 2000 mg/L at
the end of the experiment, although the simulated
concentrations of the oxidation products are somewhat
higher than those observed in column 4.
Possible reasons for the differences include inherent
assumptions of the shrinking core model (e.g. spherical
grains), an uncertain diffusion coefficient for the solid
grain cores, uncertain geochemical reaction rates,
material heterogeneity in the columns, and the limited
precision of the mineralogical analysis. There was also
b)
105
104
103
2
10
Ca 2+ (mg/L)
Total iron (mg/L)
0
Zn 2+ (mg/L)
100
200 300
Time (days)
400
500
c)
4
10
2
10
0
10
10-2
10 Column 1 suction = 2.094 m
8
6
4
Simulated
2
Mesured
0
0
100 200 300 400 500
Time (days)
0
100
200 300
Time (days)
400
400
500
c)
10
100
200 300
Time (days)
400
3
d)
400
500
e)
2
10
1
10
100
10-1
0
100
200 300
Time (days)
400
e)
1
10
100
200 300 400
Time (days)
f)
2
100
200 300 400
Time (days)
500
3
d)
1
0
0
100
200 300
Time (days)
400
500
e)
1
10
0
10-1
100
200 300
Time (days)
400
500
0
f)
1
10
0
10
-1
10
0
500
10
-1
0
500
400
10
10
500
10
10
200 300
Time (days)
10
0
10
1
10
0
10
-1
10
0
0
100
102
0
200 300
Time (days)
500
c)
0
101
100
400
b)
0
500
10
10
200 300
Time (days)
a)
-1
10
2
100
101
-1
2
10
4
10
3
10
2
10
101
100
0
10
0
d)
3
200 300
Time (days)
10 Column 4 suction = 1.30 m
8
6
4
Simulated
2
Mesured
0
0
100 200 300 400 500
Time (days)
0
500
10
100
101
10
a)
b)
4
10
3
10
2
10
101
100
0
10
0
Cu 2+ (mg/L)
a)
The differences can also be partly attributed to the
measurement time scales: while the simulated effluent
chemistry is shown for every time step (generally on the
-3
-2
order of 10 - 10 days, or ~1-15 min.), the observed
data represent an integrated response over several hours
which may not capture the rapid concentration
fluctuations during drainage.
Ca 2+ (mg/L)
SO 4 (mg/L)
pH
pH
Column1212control column
10
10 Column
88
66
44
Simulated
22
Mesured
00
0 0 100
100 200
200 300
300 400
400 500
Temps
Time (days)
(jours)
some evidence of preferential flow at the tailings/column
wall interface.
100
200 300
Time (days)
400
500
f)
1
10
0
10
10
100
200 300
Time (days)
400
500
-1
0
200
Time (days)
400
Figure 5. Simulated and observed evolution of column effluent quality (pH, sulphate and iron) for columns 12, 1 and 4.
showing a) pH, b) sulfate, c) total iron, d) calcium, e) zinc and f) copper.
1474
OttawaGeo2007/OttawaGéo2007
4.2
Sensitivity Analysis
Previous studies of EWTs have shown that to prevent
acid mine drainage, the depth of the water table must be
less than the AEV of the tailings (MEND, 1996; Ovara et
al. 1997; Dagenais, 2005). In order to further understand
the hydrogeochemical evolution of reactive tailings with
an EWT, various additional simulations were completed
using the Louvicourt tailings with the water table between
0.7 m and 6 m below the tailings surface. The simulations
were based on a large column (1.70 m) of tailings
overlayed with 0.20 cm of sand.
The simulated results shown in Figure 6 indicate that the
lower the level of the water table, the more the tailings
become desaturated. For example, for a water table at
0.7 m depth, after 396 days of simulation, the degree of
saturation of the tailings is approximately 98% compared
to only 75% at a depth of 6.2 m. The evolution of the pH,
sulphate and iron follows the trend dictated by the degree
of saturation, showing a higher degree of oxidation as the
water table elevation decreases.
For example, Figure 6 shows that after 396 days, a water
table located at a depth of 2xAEV within the Louvicourt
tailings (-6.2 m) increases the sulphate concentrations by
a factor of at least 30, and ferrous iron by a factor of 2.
Copper and zinc concentrations increase by at least a
factor of 10 compared to a water table depth of 1/5xAEV
(-0.7 m) (not shown).
1475
Elevation (m)
1.5
Water table position
-6.2 m
-4.7 m
-3.4 m
-2.6 m
-1.8 m
-0.7 m
1
0.5
0
0.2
0.4
0.6
0.8
1
Sw (%)
b)
Elevation (m)
2
1.8
-6.2 m
-4.7 m
-3.4 m
-2.6 m
-1.8 m
-0.7 m
1.6
1.4
4
6
8
pH
c)
2
Elevation (m)
The observed and simulated data show that in general,
higher suctions induced from a deeper water table (e.g.
column 1) lead to increased oxidation (lower pH, higher
SO4 & Fe) relative to the lower suction case (column4).
a)
396 days
2
1.5
-6.2 m
-4.7 m
-3.4 m
-2.6 m
-1.8 m
-0.7 m
1
0.5
0
10000
20000
30000
SO 4 (mg/L)
d)
2
Elevation (m)
The efficiency of the elevated water table method was
also assessed quantitatively from measured and
simulated concentrations of two mobile metals: copper
and zinc. Following Bussière et al. (2004) and Dagenais
(2005), the % efficiency (Ec) is defined here as Ec = (1k
k
C /Co) ×100, where C is the effluent concentration in the
test column (1 and 4) and Co is the concentration in
control column 12. The concentrations considered for the
calculations of Ec were those obtained at the end of the
simulation period (502 days) at a depth of 15 cm, i.e. at
the base of the tailings. Calculations show that the
efficiency is approximately 92% in column 1 whereas in
column 4, which had the highest water table elevation, it
reached almost 100%. These values approach those
measured by Dagenais (2005) (which lay between 95%
and 99%) who applied the EWT technique to the same
reactive tailings. The efficiencies obtained here are also
comparable with those estimated for water covers (e.g. Li
et al., 1997) and for covers with capillary barrier effects
(CCBE) (Yanful, 1993; Aubertin et al. 1999; Bussière and
al. 2004).
1.5
-6.2 m
-4.7 m
-3.4 m
-2.6 m
-1.8 m
-0.7 m
1
0.5
0
0
10
20
30
Fe 2+ (mg/L)
Figure 6. Comparison simulations at 396 days showing
the effect of various water table elevations in the tailings.
OttawaGeo2007/OttawaGéo2007
5
CONCLUSIONS
Management of reactive mine tailings is one of the most
important environmental issues facing the mining
industry. In this regard, water covers have been
recognized as one of the most effective management
techniques for controlling the production of acid mine
drainage (AMD).
The elevated water table technique (EWT) constitutes a
potentially effective alternative in which the tailings above
the water table are maintained at a nearly saturated state
by capillarity. Although this technique has recognized
advantages including a rate of oxidation comparable to
that found using water covers, a relatively low
implementation cost, and improved geotechnical stability
of the dykes, the optimal conditions and parameters are
not well defined and little lab or field data exist.
The experimental and simulated results reported herein
have shown that the hydrogeological state of the tailings
(i.e. their degree of saturation) is critical. This state is
principally controlled by the Air Entry Value (AEV). These
results confirm qualitatively that if the water table is
maintained at a depth less than the AEV of the tailings,
an EWT works well and the effectiveness can reach 99%
(similar to that of water covers and CCBEs). The
simulation methodology will allow the technique to be
optimized based on conditions that prevail at specific
mine sites.
ACKNOWLEDGEMENTS
This research was supported by the Natural Sciences and
Engineering Research Council of Canada (NSERC)
through the Industrial NSERC-Polytechnique/UQAT Chair
in Environment and Mine Waste Management, and the
Chair’s industrial and government partners. We thank Dr.
U. Mayer of the University of British Columbia for his
assistance in using the MIN3P model.
REFERENCES
Aachib, M., Mbonimpa, M. and Aubertin, M. 2004.
Measurement and prediction of the oxygen diffusion
coefficient in unsaturated media, with application to
soil covers. Water, Air and Soil Pollution, 156,
p.163-193.
Adu-Wusu, C., Yanful, E.K. and Mian, M.H. 2001. Field
evidence of resuspension in a mine tailings pond.
Can. Geotech. J., 38, p.796-808.
Amyot, G. and Vézina, S. 1997. Flooding as a
reclamation solution to an acidic tailings pound- the
th
Solbec case, Proceedings of the 4 International
Conference on Acid Mine Drainage, Vancouver,
Vol.2, p.451-464.
Aubertin, M., Aachib, M. and Authier, K. 2000. Evaluation
of diffuse gas flux through covers with GCL.
Geotextiles and Geomembranes, vol.18, p.215-233
1476
Aubertin, M., Bussière, B., Monzon, M., Joanes, A.-M.,
Gagnon, D., Barbera, J.-M., Aachib, M., Bédard, C.,
Chapuis, R.P. and Bernier, L. 1999. Étude sur les
barrières sèches construites à partir de résidus
miniers. Phase II : Essais en place. Rapport de
projet
de
recherche
CDT P1899,
École
Polytechnique de Montréal, MEND Project 2.22.2c,
331 p.
Aubertin, M., Dionne, J. and Marcoux, L. 1997. Design
guidelines and stability criteria of engineering works
th
for water covers. 4 International Conference on
Acid Rock Drainage, Vancouver, B.C., p.1849-1866
Aubertin M., Mbonimpa, M., Bussière, B. and Chapuis,
R.P. 2003. A model to predict the water retention
curve from basic geotechnical properties. Can.
Geotech. J. 40 (6): p1104-1122.
Bain, J.G., Mayer, K.U., Blowes, D.W., Frind, E.O.,
Molson, J.W., Kahnt, R., Jenk, U. 2001. Modelling
the closure-related geochemical evolution of
groundwater at a former uranium mine, J. Contam.
Hydrol., 52, p.109-135.
Bussière, B., Benzaazoua, M., Aubertin, M., and
Mbonimpa, M. 2004. A laboratory study of covers
made of low-sulfide tailings to prevent acid mine
drainage. Envir. Geol., 45 (5): p. 609-622
Catalan, L.J.J., Yanful, K.E., Boucher, J.F. and Shelp,
M.L. 2000. A field investigation of tailings
resuspension in a shallow water cover. Proceedings
from the Fifth International Conference on Acid Rock
Drainage. Vol.2, p.921-931.
David, D.-J., and Nicholson, R.V. 1995. Field
measurements for determining rates of sulphide
oxidation. Proc., Sudbury 95-Mining and the
Environment, Sudbury, Ontario, 28 May-1 June.
Dagenais A.-M. 2005. Techniques de contrôle du
drainage minier acide basées sur les effets
capillaires. Thèse de doctorat, Département de
Génie Civil, Géologique et Mines, École
Polytechnique de Montréal, 444 p.
Dagenais A.-M., Aubertin, M. and Bussière B. 2006.
Parametric study on the water content profiles and
oxidation rates in nearly saturated tailings above the
th
water table. CD-ROM, Proceedings of the 7
ICARD, Saint Louis.
Elberling B. and Nicholson R.-V. 1996. Field
determination of sulphide oxidation rates in mines
tailings, Water Resources Research, vol. 32, No.6.
p.1773-1784.
Jurjovec, J., Ptacek, C.J., Blowes, D.W. 2002. Acid
neutralization mechanisms and metal release in
mine tailings: A laboratory column experiment.
Geochemica and Cosmochimica Acta, Vol.66, No.9,
p.1511-1523.
Jurjovec, J., Blowes, D.W., Ptacek, C.J. and Mayer, K.U.
2004. Multicomponent reactive transport modeling of
acid neutralization reactions in mine tailings. Water
Resour. Res., Vol.40 W11202, 17p.
Lawrence, R.W. and Wang, Y. 1997. Determination of
neutralization potential in the prediction of acid
th
rock drainage, Proc. 4 ICARD, Vancouver, BC,
p.449-464.
Li, M., Aubé, B. and St-Arnaud, L. 1997. Considerations
in the use of shallow water covers for
OttawaGeo2007/OttawaGéo2007
th
decommissioning reactive tailings. Proc., 4 ICARD,
Vancouver, p.115-130
Mayer, K.U. 2000. MIN3P V1.0 User Guide, University of
Waterloo, Department of Earth Sciences, 26 June
2000.
Mayer, K.U., Frind, E.O. and Blowes D.W. 2002.
Multicomponent reactive transport modeling in
variably saturated porous media using a generalized
formulation for kinetically controlled reactions. Water
Resour. Res., 38, (9), 1-21 (13).
Mbonimpa, M., Aubertin, M., Aachib, M. and Bussière, B.
2003. Diffusion and consumption of oxygen in
unsaturated cover materials. Can. Geotech. J., 40:
p.916-932.
Mian, M.H. and Yanful, E.K. 2003. Tailings erosion and
resuspension in two mine tailings ponds due to
winds waves. Advances in Environmental Research,
Vol. 7, 4, p.745-765.
Molson, J.W., Aubertin, M., Bussière, B. and Joanes,
A.M. 2004. Simulating acid mine drainage through
mine wastes constructed with capillary barrier
th
covers. 57 Canadian Geotechnical Conference and
th
5
joint CGS-IAH-CNC Conference, CD-ROM,
Session 8D, p.29-36
Molson, J.W., Aubertin, M., Bussière, B. and
Benzaazoua, B. 2007. Geochemical transport
modelling of drainage from experimental mine
tailings cells covered by capillary barriers. Submitted
to Applied Geochemistry.
Ouangrawa, M., Aubertin, M., Molson,J., Zagury, G.J. and
Bussière, B. 2005. An evaluation of the elevated
water table concept using laboratory columns with
th
sulphidic tailings. 58
Canadian Geotechnical
th
Conference and 6 joint CGS-IAH-CNC Conference,
CD-ROM.
Ouangrawa, M., Molson, J., Aubertin, M., Zagury, G.J.
and Bussière, B. 2006. The effect of water table
elevation on acid mine drainage from reactive
tailings: A laboratory and numerical modeling study.
th
CD-ROM, 7 ICARD, March 26-30, 2006.
Ouangrawa, M., 2007. Analyse des facteurs d'influence
sur le comportement hydro-géochimique des rejets
miniers sulfureux partiellement submergés. École
Polytechnique de Montréal, Département de Génie
Civil, Géologique et des Mines, PhD. Thesis, in
preparation.
Orava, A.D., Tremblay, G.A., Tibble, A. and Nicholson R.
1997. Prevention of acid rock drainage through the
application of in-pit disposal and elevated water
table concepts, Fourth International Conference on
Acid Rock Drainage, Proceedings, Vol.3, Vancouver
p.973-983.
Peacey, V. and Yanful, E.K. 2003. Metal mine tailings and
sludge co-deposition in a tailings pond. Water, Air
and Soil Pollution. Vol. 145, pp. 307-339.
SRK (Steffen, Robertson and Kirsten), 1989. Draft Acid
Rock Technical Guide. BC AMD Task Force, Vol.1.
van Genuchten, M. Th. 1980. A closed-form equation for
predicting the hydraulic conductivity of unsaturated
soils. Soil Science Society of America Journal,
Vol.44, pp.892-898.
1477
van Genuchten, M. Th., Leij, F.J., and Yates, R.S. (1991).
The RETC code for quantifying the hydraulic
function of unsaturated soils. EPA/600/2-91/065.
Vigneault, B., Campbell, P.G.C., De Vitre, R. and
Prémont, S. 2001. Geochemical changes in sulfidic
mine tailings stored under shallow water. Water
Research, Vol.35, Issue 4, p.1066-1076.
Yanful, E.K. 1993. Oxygen diffusion through soil covers
on sulphidic mine tailings. ASCE, Journal of
Geotechnical Engineering. Vol.119.8; p.1207-1228.