Overview: Simulating (coupled) reactive multi-component transport with PHT3D
What is reactive transport modelling ?
Coupled/simultaneous quantification of both ...flow/transport processes
… and of reactive processes
Hydrology/Hydrogeology Geochemistry Microbiology
Interdisciplinary approach: Mainly qualitative, with few exceptions
No interdisciplinary approach
Integrated quantification of flow, transport and (bio)geochemical reactions, including isotopes
Reactive Transport Modelling: Integration of expertise from different disciplines
Why reactive transport modelling ?
Reactive transport modelling as a tool for data analysis:
• To evaluate conceptual models of reactive transport processes. • Is the conceptual model capable of explaining laboratory or field
observations ?• What are the controlling physical or chemical processes, for example
for the natural attenuation of contaminants ?• What are the parameter values (or the range of parameter values) for
reaction rate constants, dispersivities, etc., that allow to reproduce laboratory or field observations
• Hypothesis testing (Lichtner, 1996):
` Quantitative models force the investigator to valid ate or invalidate ideas by putting real numbers into an of ten vague hypothesis and thereby starting the thought p rocess along a path that may result in acceptance, rejecti on, or modification of the original hypothesis'.
Why reactive transport modelling ?
Reactive transport modelling as predictive tool:• To what extent will environmentally important receptors downgradient
of the source zone be impacted by a contaminant ?• What are maximum concentration levels and what is the contaminant
mass flux ?• What are the time-scales for cleanup to below given limits for different
remediation schemes ?• What is the optimal design of a particular (active/passive) remediation
scheme ?
However, predictions are strongly effected by uncertainty that originates from:
• Incomplete hydrogeological and hydrogeochemical site characterization
• Incomplete or completely unknown ‘source’ history• Incomplete process understanding, wrong conceptual models
Complexation of elements Ca2+ + SO4
2- = CaSO4
Weathering of minerals KAlSi3O8 + 8 H2O = K+ + Al(OH)4
- + 3 H4SiO4
(Dissolution of K-feldspar)
Precipitation of mineralsFe2+ + HS- = FeS + H+
(Precipitation of amorphous iron-sulphide - FeS(ppt))
Reactive processes (Examples)
Degradation/mineralisation of dissolved and sediment-bound organic matter
CH2O + O2 → HCO3- + H+
(Aerobic degradation of organic matter)
Reactive processes (Examples)
Degradation/mineralisation of organic contaminants
C7H8+ 21H2O ⇒ 7 HCO3- + 43H+ + 36e-
(Mineralisation of toluene)
Transformation of organic contaminants
C2H4Cl2 ⇒ C2H3Cl + Cl- + H+
(Transformation from Dichlorethane to Vinylchloride)
Reactive processes (Examples)
Microbial growth5CH2O + 0.6NH4
+ + 2O2
→ 0.6C5H7O2N + 2HCO3- + 2.6H+ + 1.8H2O
(Degradation of organic matter and incorporation of organiccarbon into biomass)
Reactive processes (Examples)
Complexation with ion-exchanger site Na+ + X- = NaXK+ + X- = KXLi+ + X- = LiXCa+2 + 2X- = CaX2
X (Exchanger site)
KK Li K NaNa Ca
→ Number of sites on the ion-exchanger X is limited
→ Sites are all occupied
→ Sites are provided e.g., by mineral surfaces
Reactive processes (Examples)
Complexation of metals with charged surfaces
Hfo_wOH + Cd+2 = Hfo_wOCd+ + H+
(Sorption of Cadmium to Hydrous ferric oxide - Hfo)Hfo_wOH + H+ = Hfo_wOH2+
(Sorption of Hydrogen to Hydrous ferric oxide - Hfo)
→ Cations and anions complex with mineral surfaces
→ Hydrogen on surface sites
→ Results in pH-dependent mobility of trace metals
Reactive processes (Examples)
Time-scales of reactive processes
Damköhler number
i
ii vc
xDa
,0
ϕ=
Da → 0 Tracer transport (No reaction)
Da → ∞ Local equilibrium assumption (LEA) valid
size of grid-cell
flow velocity
reaction rate
reference concentration
(Domenico and Schwartz, 1990)
LEA invalidLEA valid
Kinetically controlled reactionsEquilibrium reactions
Time-scales of reactive processes
Discretization/scale dependency of LEA
(1) Column experiment, residence time = 1day
Dissolution of K-feldspar (LEA not valid)
Dissolution of Calcite (LEA valid)
(2) Regional aquifer, residence time in grid-cell > 1 year
Dissolution of K-feldspar (LEA valid ?)
Dissolution of Calcite (LEA valid)
Single-species transport modelsReaction rate does not depend on the concentration of other species
Multi-species transport modelsFate of multiple chemicals is simulated
Reaction rates can depend on concentrations of (many) other species
Multi-component transport modelsSimulation of “full” geochemistry, including pH,
redox-state, sediment/rock – water interaction
Classes of reactive transport models
Distance from the Source
Concentration
Species 1
Constant-concentration source containing species 1 only
Single-species transport models
Distance from the Source
Concentration
Species 1
Species 2
Species 3
Constant-concentration source containing species 1 only
Multi-species transport models
Multi-species and multi-component models
RT3D (Clement, 1997)MT3D99 (Zheng/SSPA, 1999)BIOREDOX (Carey et al., 1999)SEAM3D (Waddill and Widdowson, 1998)BIONAPL (Molson, 2002)
CRUNCHFLOW (http:www.csteefel.com)PFLOTRAN (http://ees.lanl.gov/pflotran)STOMP (www.stomp.pnl.gov) PHT3D (www.pht3d.org)PHAST (Parkhurst et al., 1995)MIN3P (Mayer, 1999)HBGC123D (Salvage and Yeh, 1998)
… focussing on ‘primary biodegradation reactions’
… with comprehensive geochemical capabilities
Some available simulators:
Reactive Multi-Component Transport Model PHT3D
PHT3D
CSIRO/UWA/Flinders Uni
MT3DMS
Univ. of Alabama
PHREEQC-2
USGS
Advective-dispersive transport
MODFLOW
USGS
Groundwater flow
Model applications
Toolbox for 3-D reactive multicomponent transport: Definition of reaction modules
Contaminated sites MAR sites
Visual Modflow, Aquachem
WHI, Schlumberger
Prommer et al., 2003. Ground Water; Post and Prommer, 2007 , WRR
Processing Modflow
Simcore
Graphical User Interfaces
Mining (ISL)
Geochemical processes
Biodegradation model with geochemical capabilities was needed
Initial Motivation for PHT3D development
• Thin BTEX plumes, indicating small dispersivity• Sulphate depletion but no sulphide observed
• Fe(III) acting as electron acceptor ?(precipitation of pyrite, siderite, magnetite ?)
• Three-dimensional observation data from multiportsavailable
(dissolution of ferrihydrite, goethite ?)
Initial solution (1998)
Current solution, PHT3D v2.14
• MT3DMS (incl. div. MOC + TVD schemes)
• Geochemical equilibrium model PHREEQC (v1.6) Biodegradation module (Monod kinetics)
• NAPL dissolution module (kinetic, Raoult’s law)
• Coupling via sequential, non-iterative operator-splitting
• MT3DMS (v5.3)
• PHREEQC-2, mixed equilibrium and kinetic reactions (v2.17)
• All kinetic reactions are formulated (and easily modified) in the PHREEQC-2 database, including microbial reaction, NAPL dissolution, etc.
PHT3D
PHT3D
MT3DMS step, advective-dispersive transport • All MT3DMS features available• Transport of components (SOLUTION_MASTER_SPECIES)
PHREEQC-2 step, reactions• Aqueous complexation/speciation (equilibrium)• Kinetic reactions of aqueous species/components
(e.g., biodegradation of organic compounds)• Mineral precipitation/dissolution (equilibrium/kinetic)• Ion-exchange (equilibrium)• Surface complexation model (SCM) since v2.0• Dual-domain
•
Multi-component vs Multi-species transport
Transport of total aqueous component concentration computed by MT3DMS
Source/sink term for all reactive processes (precip./dissolution,
mineralisation of organic compounds, sorption, etc), computed by PHREEQC-2
Operator Splitting (OS)
• (Temporal) operator-splitting method: Separate solutions for the “transport step” and “reaction step(s)”
• Rk’s are independent from neighbouring grid cells and might be computed in parallel mode. Rk’s are independent from neighbouring grid cells
• Splitting introduces “error”
• Iteration between the steps would reduce this “error”
MT3DMS PHREEQC
PHREEQC runs a reaction simulation for each grid cell
MT3DMS simulates the transport of n components for a timestep ∆t
Comp 1 Comp 2 … Comp n
Operator Splitting
PHREEQC runs a reaction simulation for each grid cell
… Comp n
MT3DMS simulates the transport of n components for a timestep ∆t
Comp 2Comp 1
etc …
Operator Splitting
PHT3D Implementation of Operator splitting (III)
Time level k-1 k+1k
∆t ∆tTime step length
Transport step length ∆ttr1 ∆ttr1
∆t
∆ttr2 ∆ttr1 ∆ttr1 ∆ttr2 ∆ttr1 ∆ttr1 ∆ttr2
∆t ∆tPHREEQC-2 step length ∆t
Integration interval length ∆tr3∆tr1 ∆tr2 ∆tr1 ∆tr2 ∆tr1
∆ttr1 ∆ttr1 ∆ttr2 ∆ttr1 ∆ttr1 ∆ttr2 ∆ttr1 ∆ttr1 ∆ttr2PHT3D (“safe” option)
∆t ∆t∆tPHT3D (default option)
MODFLOW
bas.dat
bcf.dat
drn.dat
pcg2.dat
…
PHT3D
pht3dbtn.dat
pht3dadv.dat
pht3ddsp.dat
pht3dssm.dat
budget.dat
heads.dat
ddown.dat
…mt3d. flo
PHT3D001.UCN
...
PHT3Dxxx.UCN
pht3d_ph.dat
pht3d_datab.dat
GUI
GUI
PHT3D: Structure and Data Flow
pht3dgcg.dat
Mineral Precipitation/Dissolution• Migration of precipitation/dissolution fronts (Engesgaard &
Kipp, 1992) • Acid mine drainage/mineral buffering (Walter et al., 1994)
Ion Exchange• Flushing of a Na-K-NO3 -solution with Ca-Cl2 (Appelo, 1994)• Artificial recharge (Valocchi, 1981)• Anionic tenside injection (Sardin et al., 1986)
Kinetic Reactions• Single-species biodeg./Monod kinetics (Parlange, 1984)• Sequential/parallel decay chain (Sun et al., 1999) • Hydrocarbon degradation using multiple electron acceptors,
RT3D reaction module 3 (Clement, 1997)• Sequential degradation of CHCs,
RT3D reaction module 6 (Clement, 1997)
Verification / benchmark problems
Ion Exchange – Valocchi et al. (1986)
102
103
104
105
10−4
10−3
10−2
10−1
100
Observation well S 23
C (
mol
/l)
Mg (PHREEQC−2)Ca (PHREEQC−2)Na (PHREEQC−2)Mg (PHT3D) Ca (PHT3D) Na (PHT3D) Mg (observed) Ca (observed)
0 200 400 600 800 1000 12000
0.05
0.1
0.15
Volume injected (m3)
C (
mol
/l)
Cl− (PHREEQC−2)Cl− (PHT3D) Cl− (observed)
Attenuation of an Ammoniacal Liquor Contamination at a former Coking Plant in the UK
-10
0
10
20
30
40
50
60
70
80
90
0.0E+00 1.3E-02 2.5E-02
0.0E+00 1.3E-02 2.5E-02
-10
0
10
20
30
40
50
60
70
80
90
0.0E+00 1.3E-02 2.5E-02
0.0E+00 1.3E-02 2.5E-02
-10
0
10
20
30
40
50
60
70
80
90
0.0E+00 1.3E-02 2.5E-02
mA
OD
0.0E+00 1.3E-02 2.5E-02
Haerens, B. (2004). Reactive transport modelling of a groundwater contamination by ammoniacal liquor at a former coking plant near Rexco/UK. PhD thesis, KU Leuven, Belgium
Applications: Biodegradation of a Hydrocarbon Plume under Transient Groundwater Flow Conditions
Prommer, H., Davis, G.B., and Barry, D.A. (2002). Modelling of physical and reactive processes during biodegradation of a hydrocarbon plume under transient groundwater flow conditions. J. Cont. Hydrol. 59, 113-131.
Understanding Contaminant Degradation Behaviourthrough Measuring and Modelling of Isotope Signatures
Prommer., H., Anneser, B., Rolle, M., Einsiedl, F., Griebler, C. (2009). Biogeochemical and isotopic gradients in a BTEX/PAH contaminant plume: model-based interpretation of a high-resolution field data set. Environ. Sci. Technol.
Source zonesimulatedmeasured
mol/L
Source zone
+4‰
‰
Fate of specific Polycyclic Aromatic Hydrocarbons in a complex Mixture of Organic Compounds
D’Affonseca, F.M., Prommer, H., Finkel, M., Blum, P., and Grathwohl, P. (2010). Long-term evolution of transient biogeochemical and isotopic signatures in a coal tar contaminated aquifer. Water Resour. Res.
Modelled2D transect
Redox Zonation during Injection of Surface Water into an anoxic Aquifer
0 100 200 300 400 500 600 700 8000
123
456
x 10-4
WP1-F3
N(5
) (m
ol/L
)
Time (days)
T= variable
T= 14oC
T= 20oC
T= 8 oC
Prommer and Stuyfzand (2005). Prommer, H. and Stuyfzand, P. J. (2005). Identification of temperature-dependent water quality changes during a deep will injection experiment in a pyritic aquifer, Environ. Sci. Technol., 39(7), 2200-2209.
Impact of Redox Zonation on the Fate of Pharmaceuticals during Artificial Recharge
Greskowiak, J., Prommer, H., Massmann, G., and Nützmann, G. (2006). Modeling seasonal redox dynamics and the corresponding fate of the pharmaceutical residue phenazoneduring artificial recharge of groundwater Environ. Sci. Technol., 40 (21), 6615–6621.
0
10
20
30Temperature
°C
0
2
4
x 10-4
Dissolved oxygen
mol
/l
0 200 400 600 800 10000
0.2
0.4
0.6
days
µgl/l Non-reactive simulation
Reactive simulation
Phenazone
Application: Quantifying Biogeochemical Changes during Aquifer Storage and Recovery of Reclaimed Water
Greskowiak, J.; Prommer, H.; Vanderzalm, J.; Pavelic, P.; Dillon, P. (2005). Modelling of carbon cycling and biogeochemical changes during a wastewater injection and recovery experiment at Bolivar/South Australia. Water Resour. Res. 41, doi:10.1029/2005WR004095
0
0.005
0.01
0.015m
eq/l
ASR well
Alkalinity
0 200 400 600 8006
6.5
7
7.5
8
Days since start of injection
pH
Layer 350m well
ASR well
200m
Quantifying Arsenic Fate during Managed Aquifer Recharge
Wallis, I., Prommer, H., Pichler, T., Post, V.E.A., Norton, S., Annable., M. and Simmons, C.T. (2011). A process-based reactive transport model to quantify arsenic mobility during aquifer storage and recovery of potable water. Environ. Sci. Technol.
ASR well Observation well
Fate of Phenoxy Acids in a Landfill Leachate Plume
Prommer, H., Tuxen, N., and Bjerg, P. (2006). Fringe-controlled natural attenuation of phenoxy acids in a landfill plume: Integration of field-scale processes by reactive-transport modelling. Environ. Sci. Technol., 40, 4732-4738.
Uranium transport at the Hanford 300A site
meters
Multi-rate surface complexation model
Equilibrium surface complexation model
U(VI)
U(VI)
Ele
vatio
nE
leva
tion
Ma, R., C. Zheng, H. Prommer, J. Greskowiak, C. Liu, J. Zachara, and M. Rockhold (2010), A field-scale reactive transport model for U(VI) migration influenced by coupled multirate mass transfer and surface complexation reactions, Water Resour. Res., 46
Application: Modelling in situ Bioprecipitation of Zn and Cu: Column Experiments
Prommer, H., Grassi, M.E., Alexander C. Davis, A.C., and Patterson, B.M. (2007)., Modeling the effects of increasing acidity on bacterial sulfate reduction and on metal bioprecipitation in groundwater affected by acid mine drainage. Environ. Sci. Technol., 41(24), 8433-8438.
• Addition of degradable organic compounds to stimulate sulfate reducing conditions
• 2CH2O + SO4-2
→→→→ 2 HCO3- + H2S
• Me2+ + H2S
→→→→ MeS(solid) + 2 H+
• Longevity of metal immobilsation ?
• Impact of acidity on remobilisation ?
Measu
red
Modelled
Application: Feedback Mechanism between Variable-Density Flow and Geochemical Reactions
Post, V.E.A., and Prommer, H. (2007). Reactive multicomponent transport simulation of the Elder problem: effects of chemical reactions on salt plume development. Wat. Resour. Res. 43 (10)
Bauer-Gottwein, P., Langer, T. Prommer, H., Wolski, P. and Kinzelbach, W. (2007). Okavango Delta Islands: Interaction between density-driven flow and geochemical reactions under evapo-concentration. J. Hydrol., 335, 389– 405.
Application: Reactive Transport under Variable-Density Conditions: Okavango Delta Islands
50 100 150 200-50
-40
-30
-20
-10
50 100 150 200-50
-40
-30
-20
-10
50 100 150 200-50
-40
-30
-20
-10
50 100 150 200-50
-40
-30
-20
-10
-4
-3
-2
-3
-2
-1
-5
-4
-3
7.5
8
8.5
9
9.5
50 100 150 200
-4
-2
50 100 150 200-3
-2
-1
50 100 150 200
-5-4-3-2
50 100 150 2007
8
9
Cl
pHCa
Alk
Spatial scale of subsurface transport problemChemical complexity
Batch (0D) Column (1D) Field (2D/3D)
1 10 100 1000 10000 m
PHT3D
Rexco
Sjoelund
Fe0-column
Model-complexity and spatial scale of typical PHT3D applications
Ethanol columns
Dizon
NIT
Okavango Delta Islands
Bolivar ASR
Single/multi-species transport models
PHREEQC-2
Warning:Reactive transport ≠ regional MODFLOW model + reactions
• Vertical chemical gradients may be much more significant than lateral chemical gradients
• Vertical grid resolution of flow models is often insufficient toresolve these vertical chemical gradients – especially were vertical transverse mixing controls the progress of reactions
• Even pefectly calibrated flow models may not predict transport behaviour very well – chemical data needed as additional constraints
• (Desktop) computational power is only slowly sufficient to allow 3D (high resolution) models
Warning:Reactive transport ≠ regional MODFLOW model + reactions
Be smart• Think hard(er) about the combined conceptual
hydrogeological + chemical model• Is perhaps a 2D vertical transect sufficient to address the
reactive transport problem ?• Is a 3D reactive transport model for a subregion of the flow
model sufficent ?• Is the problem symmetric ? (ASR, Deepwell injection, ....)• Can a rough calibration be carried out on a less refined grid
and grid resolution successively increased ?
From simple to complex• Test + debug chemical reactions (e.g, rate expressions for
kinetics) in 0D (PHREEQC batch mode).• Build a mickey mouse model with reactions (1D, 2D) and
debug problems + get a feel for the processes
Initial and boundary conditions• Formulation of initial and boundary conditions is similar to
MT3DMS. • Note, that the same type of boundary condition applies to all
components/compounds.• Water chemistry (solution composition) must be defined for all
boundaries with a positive flux into the model, otherwise “de-ionized” water is added to the domain (C = 0 for all species)
• Solution compositions at boundaries must be charge-balanced and (depending on the conceptual model) be pre-equilibrated with the minerals
• Initial solution composition(s) should be charge-balanced and pre-equilibrated with the mineral assemblage etc.
PHT3D: Definition of initial and boundary conditions
(2) Recharge water composition
(3)
(2) (1)
(4) (Charge-balanced/equilibrated(?)) water composition
(4) (5)
(3) (Charge-balanced/equilibrated(?)) inflow across boundary
(1) (Charge-balanced/equilibrated) initial water composition + minerals + X …(2) Outflow across boundary
(5) Extraction water composition
FD+Fast, mass-conservative- Numerical dispersion in case of advection-dominated transport
MMOC+Reliable in many cases- Numerical dispersion in case of advection-dominated transport
TVD+Little or no numerical dispersion in case of advection-dominated
transport+Faster than MOC-schemes - Some difficulties if water table is crossing multiple layers
MOC, HMOC+Little or no numerical dispersion in case of advection-dominated
transport- Needs to be used in conjunction with a “slower” PHT3D
coupling scheme
PHT3D: Considerations for the MT3DMS advection scheme
PHT3D: Visualisation and Postprocessing
Pre/postprocessing• Visual Modflow (> v4.1) • Processing Modflow (>= v8)• Groundwater Vistas (in preparation)• ipht3d (phython-based free tool, in preparation)
3D Visualisation options• 3D-Master• Model Viewer (USGS, free)
Advanced, more scientific postprocessing• Mass balance, plotting integrated mass vs time, etc) requires tools
such as MATLAB or python• ASCII-format can be processed by MATLAB, TECPLOT, etc• Many MATLAB scripts to postprocess data have already been
produced over the years. They can be adapted with some minor effort
Definition of Initial Conditions/Concentrations
Case 1: Geochemical equilibrium is a good approximation for the definition of the initial concentrations
• SI can be <= 0 (under-saturated) for minerals with C0 = 0 • If the SI of an equilibrium mineral is > 0 (over-saturation) then
precipitation will occur during the initial calibration of the model (and then perhaps lead to unexpected modifications of the solution composition)
• If the SI of an equilibrium mineral is < 0 (under-saturation) and C0 > 0 then dissolution will occur during the initial calibration of the model (and then perhaps lead to unexpected modifications of the aqueous initial composition)
• The ion exchanger occupation is in equilibrium with the aqueous solution
• The surface site occupation is in equilibrium with the aqueous solution
Definition of Initial Conditions/Concentrations
Case 2: Quasi-dynamic equilibrium • Slow, kinetically controlled weathering reactions … SI does not
or only very slowly reach 0 within the model domain …. • Example: Kinetically controlled DOC or SOM degradation or
other processes cause a redox zonation within aquifer
→ Apply spin-up period
... e.g., model domain completely flushed once