Modeling AMD
Geochemistry in
Underground Mines
Bruce Leavitt PE PG, Consulting Hydrogeologist
James Stiles PhD PE, Limestone Engineering
Raymond Lovett PhD, Shipshaper LLC
Limitations of existing AMD
Prediction Methods
Only
considers Acid and Base Potential
Does not consider Latent Acidity
Does not consider Oxygen Depletion
Does not consider Solute Transport
Does not consider Recharge Water
Chemistry and Volume
Study Purpose

To investigate the suitability of the model
to underground mine discharges.
 To determine the appropriate mineral
assemblage and mass concentration.
 To compare the model in different
hydrologic settings.
 To evaluate the sensitivity of the model to
variations in input values comparable to
typical field variations.
Three Hydrologic Settings
Unflooded, Free Draining
overburden
Mine
Discharge
River
Flooded Mine Low Dilution
Flooded High Dilution
Mine
Pump
overburden
Mine
Discharge
River
No Discharge
overburden
River
Effect of Flooding on Mine
Water Chemistry

Rapid dissolution of acidic salts
 Exclusion of oxygen from the mine
 Chemical reaction with recharging
ground water.
TOUGHREACT
Earth Sciences Division, Lawrence Berkeley
National Laboratory
TOUGHREACT was designed to solve the coupled
equations of sub-surface multi-phase fluid and heat flow,
solute transport, and chemical reactions in both the
saturated and unsaturated aquifer zones. This program can
be applied to many geologic systems and environmental
problems, including geothermal systems, diagenetic and
weathering processes, subsurface waste disposal, acid mine
drainage remediation, contaminant transport, and
groundwater quality.
Model Configuration
Mineral Assemblage
Volume
Concentration
K25 (mol/m2/s)
Ea (kJ/mol)
calcite
0.001
equilibrium
equilibrium
gypsum
0.0001
equilibrium
equilibrium
melanterite
0.002
equilibrium
equilibrium
rhodochrosite
0.010
3.55x10-6
40.0
illite
0.400
6.9185x10-13
22.2
jarosite
0.001
6.9185x10-13
22.2
Al(OH)3
(amorphous)
0.001
6.9185x10-13
22.2
gibbsite
0.001
6.9185x10-13
22.2
pyrolusite
0.001
6.9185x10-13
22.2
Mineral
Mineral Assemblage cont.
Volume
Concentration
K25 (mol/m2/s)
Ea (kJ/mol)
ferrihydrite
0.001
6.9185x10-13
22.2
jurbanite
0.001
1.0233x10-14
87.7
quartz
0.001
1.0233x10-14
87.7
Neutral 6.918x10-14
Acid
4.898x10-12
Base
8.913x10-18
Neutral 3.020x10-13
Acid
7.762x10-12
Base
N/A
Neutral 2.818x10-6
Acid
3.020x10-9
Base
N/A
Neutral 1.660x109-9
Acid
2.570x10-4
Base
N/A
Neutral 1.260x109-11
Acid
6.457x10-9
Base
N/A
22.2
65.9
17.9
88.0
88.0
N/A
56.9
56.9
N/A
62.76
36.1
N/A
18.6
18.6
N/A
Mineral
kaolinite
0.500
chlorite
0.001
pyrite
0.0015
siderite
0.001
magnetite
0.001
Archetype pH
Observed Water Chemistry
pH
9
8
Free Draining
Standard Units
7
High Dilution
6
Low Dilution
5
4
3
2
0
5
10
15
20
Years Since Mine Flooding
25
30
Archetype Iron
Observed Water Chemistry
Iron
1200
1000
Free Draining
mg /L
800
High Dilution
Low Dilution
600
400
200
0
0
5
10
15
20
Years Since Mine Flooding
25
30
Model Results pH
pH
Free Draining
Low Dilution
High Dilution
Flooding Time
8
7
6
pH
5
4
3
2
1
0
4
8
12
Simulation Time, years
16
20
Model Results Iron
2000
Total Iron
Free Draining
Low Dilution
High Dilution
Fill Time
Total Iron, mg/L
1600
1200
800
400
0
0
4
8
12
Simulation Time, years
16
20
Pyrite Kinetic Data

Neutral 2.818 x 10-6 mol-m-2-s-1 McKibben and
Barnes (1986a)

Neutral 3.167 x 10-10 mol-m-2-s-1 McKibben and
Barnes (1986b), Nicholson (1994), and Nicholson and Sharer (1994)
Acidic 3.020 x 10-9 mol-m-2-s-1
 Acidic 1.553 x 10-8 mol-m-2-s-1 McKibben and Barnes

(1986b), Brown and Jurinak (1989), and Rimstidt, et al. (1994)

Acidic 6.0 x 10-10 mol-m-2-s-1 Calibrated
Ferrous Ferric Oxidation


Fe+2 + 1/4O2 + H+ > Fe+3 +1/2 H2O
Oxidation rate is pH dependant.
 Model holds ferrous and ferric iron in
equilibrium.
 Model overstates ferric iron concentration
leading to excess pyrite oxidation.
High Dilution pH
Year 5
High Dilution pH
Year 10
High Dilution pH
Year 15
High Dilution pH
Year 20
High Dilution Iron
Year 5
High Dilution Iron
Year 10
High Dilution Iron
Year 15
High Dilution Iron
Year 20
Modeling Difficulties

Ferrous iron oxidation
 Insufficient aluminum production
 CO2 partial pressure spikes at full mine
flooding
 Mine complexity is limited by
computational capacity
 Homogeneous mineral distribution
 Mine atmosphere composition
Other Results

Gypsum precipitation / dissolution in the
mine
 Goethite precipitation in the mine.
 Elimination of pryhotite and the reduction
of the pyrite kinetic rate has reduced the
observed difference in water pH and iron
between the high dilution and low dilution
cases.
Future Work

Resolve the iron oxidation issue
 Closed mine atmosphere sampling.
 Sensitivity analysis of input parameters
including: recharge chemistry, mine
geometry, initial melanterite and calcite
concentrations.
 Testing of in situ remedial options.
Conclusions

The TOUGHREACT program allows chemical
and hydrodynamic interaction in a flooded and
unflooded underground mine environment.
 TOUGHREACT is able to emulate the change in
discharge chemistry with time.
 It is a useful tool in understanding acid formation,
solute transport, and discharge relationships.
 Due to the extensive number of assumptions it is
not, at this time, a suitable permitting tool.