Contaminant Transport — Plume Concentration Modeling

12.1 Transport Auto-Activates

Unlike most groundwater modeling platforms, IGW-NET has no separate on/off checkbox for contaminant transport. The moment you assign a non-zero concentration anywhere — a source zone polygon, a zone override, a well's injection concentration, a polluted stream's concentration, the recharge concentration field — transport simulation activates automatically and the solver begins tracking the resulting plume.

12.1.1 Why auto-detection matters

This design choice reflects a practical observation about how modelers actually work. If transport is a checkbox, users routinely forget to turn it on after defining a source, and then wonder why the plume isn't appearing. Conversely, leaving transport "on" with no sources wastes computation on a trivial no-transport solution. Auto-detection eliminates both failure modes: no source means no transport simulation; any source means transport is simulating. The user experience matches the physics — transport happens if and only if there's something to transport.

12.1.2 How to tell if transport is active

Three indicators confirm transport is running:

• The Simulate button's behavior — if transport is active, the simulation takes modestly longer than flow-only and the resulting animation includes plume migration.
• The Analysis Tools — concentration charts, breakthrough curves, and the transport mass balance become available only when transport is active.
• The automatically-generated model description (via the Intelligent Reporting System) — explicitly states "Contaminant transport is active" or "No contaminant transport," with the list of detected sources.

12.2 Where Contamination Comes From

The single most important concept in this chapter: contamination can enter your model from many distinct pathways, and getting the right pathway right is often more consequential than getting the physical processes right. This section presents the catalog in two voices — the applied-practitioner voice (how an EPA manager, consultant, or site investigator thinks about contamination) and the modeler voice (how it maps to IGW-NET's boundary-condition controls).

12.2.1 The applied view — contamination as it actually happens

Real-world contamination enters aquifers through recognizable pathways. A site investigator or EPA manager doesn't think in terms of boundary conditions; they think in terms of what actually happened at a site.

12.2.2 The modeler view — the same pathways as boundary conditions

Inside the simulator, every source in the applied list above is expressed as one of five modeler-recognizable configurations. The vocabulary is the same you'll see in IGW-NET's dialogs, in published MODFLOW/MT3DMS documentation, and in any groundwater-modeling textbook.

12.2.3 The rule, in both voices

The single rule that ties everything together — expressed twice, once for each audience — determines whether any given feature can be a source.

12.2.4 Features that are categorically never sources

This list exists to prevent a common mistake. The following features cannot be contamination sources, regardless of how they are configured:

If you find yourself trying to introduce contamination through one of these features, you need a different feature. A polluted drainage ditch that represents contaminated water recharging the aquifer should be modeled as a two-way stream or as a head-dependent polygon, not as a one-way drain. An extraction well that's part of a recirculation system where treated water goes back in should be two wells (extraction + injection), not one reversible well.

12.3 Setting Up Sources

This section walks through the mechanics of defining each source type. The applied and modeler categories from §12.2 map to specific dialogs and controls in IGW-NET.

12.3.1 Areal source zones — continuous and instantaneous

The most common source configuration in practice: an area of the domain where contamination is released, either continuously or as a pulse at t = 0.

12.3.2 Zone concentration overrides — initial conditions within a zone

Distinct from the source-zone release pattern above, a zone can also have a concentration override that sets the aquifer's initial concentration within that zone. This is how you represent a pre-existing plume at t = 0 — a legacy plume you're starting to model with some characterization already known.

In the Zone Attributes dialog, the concentration field assigns an initial value at every cell inside the zone. If you specify a source polygon for ongoing release plus a zone override for initial concentration, you get both effects: contamination already present at t = 0, plus continued release from the source.

12.3.3 Wells with injection concentration

An injection well with a non-zero concentration adds mass at the well cell each time step. Chapter 6 §6.4 introduced the well record with its 13 fields; for transport sources, the relevant field is Conc_ppm:

• Set Q > 0 (positive, meaning injection).
• Set Conc_ppm to the concentration of the injected water.
• The solver adds mass at a rate of Q × Conc_ppm at the well cell, each time step.

For an injection well with a time-varying pumping schedule (Ch. 9 §9.4), the injection concentration can likewise be time-varying — e.g., modeling an in-situ remediation system where reagent concentration changes with operating procedures.

12.3.4 Polluted rain recharge

Recharge always enters the aquifer from above, so a non-zero concentration assigned to the recharge field represents contamination dissolved in infiltrating precipitation. This handles:

• Agricultural fertilizer and pesticide leaching — often spatially distributed by land use (assign via zone overrides to give different recharge concentrations to different land-use categories)
• Atmospheric deposition from nearby sources — typically spatially uniform unless concentrated dispersion patterns are modeled
• Acidic or chemically-altered recharge from surface disturbance
• Leaking septic systems operating as areally distributed surface-layer sources

Where recharge concentration varies spatially (land-use patterns, distance from a surface source), use zones or uploaded rasters to carry the spatial pattern.

12.3.5 Polluted streams and lakes (two-way features)

For a polluted stream or lake to contribute as a source, three conditions must hold: (1) the feature is configured as two-way head-dependent (not one-way drain); (2) the feature has a non-zero concentration in its attributes; (3) the computed flow direction is into the aquifer (the feature is losing water at that segment).

In the Edit Polyline Attributes dialog for a two-way stream (type 5), or the Zone Attributes dialog for a two-way head-dependent lake polygon, you will find a concentration field. Fill it with the stream or lake water's contamination level. The platform handles the rest — during time steps when the feature is losing, contamination enters the aquifer; during time steps when the feature is gaining, no mass is added (the aquifer's own concentration is carried out instead).

12.3.6 Prescribed-head, prescribed-flux, and transient-head boundaries

Any boundary segment where water enters the aquifer can carry concentration. For prescribed-head polylines (types 1, 2, 3) and the head-from-transient-file polyline (type 8), the dialog includes a concentration field that applies to any inflowing water. For prescribed-flux polylines (type 7), the concentration applies to the specified inflow flux.

The platform handles the direction test automatically — it will apply the concentration only at segments and time steps where water is flowing in, and never at segments or times where water is flowing out.

12.4 The Five Physical Processes — and How to Invoke Them in IGW-NET

Once you've defined your sources, five physical processes govern how the resulting plume evolves. Each has a distinct physical mechanism; each is controlled by specific fields in IGW-NET with specific defaults. This section walks through each process as the platform handles it: what you do to turn it on, what you see when you do, and the IGW-NET-specific judgment calls that decide whether your parameter value is right for your problem.

Two organizing facts that recur for every process:

• Transport is auto-activated by concentration sources (§12.1), but each individual process must be enabled separately. The platform's default transport simulation is advection-only. Dispersion, diffusion, sorption, and decay are opt-in — you activate each by checking its flag in the Aquifer Attributes dialog and supplying parameters.
• When a process is off in IGW-NET, its parameters don't matter. The platform ignores αL if the dispersion flag is off, ignores Kd if the sorption flag is off, and so on. Setting non-zero values without checking the activation flag is a common "why isn't this working?" source of confusion.

12.4.1 Advection — always on

Advection is the only one of the five processes that is always active in IGW-NET. The platform transports dissolved mass with the local groundwater velocity without any flag to enable it and without any parameter to tune — it is the baseline that the other four modify. This matches the physics: you cannot have dissolved transport without the bulk flow of water carrying it.

Because advection velocity comes directly from the flow solution, your transport results are only as good as your flow solution. If your hydraulic conductivity is wrong, plume arrival times will be wrong proportionally. If your regional gradient is off, your plume goes in the wrong direction. For this reason, the standard IGW-NET transport workflow is: calibrate flow first (Ch. 17), verify heads match observations, then add transport. Trying to fit transport data with a poorly calibrated flow field burns parameters on compensating for flow errors.

12.4.2 Mechanical dispersion — the Dispersivity dialog

Dispersion represents plume spreading from velocity variations smaller than the grid resolves: pore-scale variability, intra-cell heterogeneity, velocity differences between streamlines. In IGW-NET, you invoke it through the Dispersivity controls inside the Aquifer Attributes dialog. The field carries four values internally — an activation flag and three dispersivity coefficients.

The platform's default of zero-and-off is deliberate: a new user's first model gives them a pure-advection plume that is easy to interpret visually, and they add dispersion only when they decide their problem calls for it. The visual consequence of turning dispersion on: the plume becomes blurrier at its edges, concentrations at the leading edge drop below the source value, and the plume envelope grows with the square root of travel distance rather than staying as a sharp advection front.

Choosing a value for α_L: the practical approach in IGW-NET is to pick a value appropriate for your plume's spatial scale, simulate, and compare the simulated plume's spread to whatever observational data you have. A rule of thumb: α_L is typically a few percent of plume length. If your plume is 500 m long, α_L in the 5-50 m range is reasonable; if the simulated plume looks too sharp compared to observations, increase α_L; if too diffuse, decrease.

12.4.3 Numerical dispersion — when the grid spreads the plume

After you set up physical dispersion with α_L, α_T, α_V, your simulated plume will always spread somewhat more than those values alone predict. The additional spreading is numerical dispersion, and it arises from the fundamental way finite-difference transport solvers represent concentration: as a single average value per cell.

Why it happens — the averaging effect

Concentration inside a real plume varies continuously in space. A finite-difference solver must collapse that continuous distribution into one number per cell — the cell-averaged concentration. When a cell is large relative to the plume's features, averaging over the cell smooths out concentration gradients that really exist on smaller scales. Every time the solver advances contamination across a cell boundary, it re-averages the receiving cell's contents; each averaging step blurs the plume a little more. The cumulative effect over many cells and many time steps is plume spreading that looks like dispersion but comes from the grid, not the physics.

Two consequences are worth understanding because they point in opposite directions:

• Plume extent is slightly over-predicted. Numerical dispersion pushes the plume's leading edge a bit further than physical dispersion alone would, because mass has been smeared ahead of the true front. In a regulatory context, this is conservative — the simulated plume reaches receptors earlier and spreads wider than reality. If your question is "does the plume reach this receptor?", a coarse-grid simulation is likely to say yes more often than reality.
• Peak concentration is under-predicted. The same averaging that smears the leading edge also dilutes the plume core — the same total mass is distributed across a wider and flatter plume, so the peak value is lower than it should be. If your question is "does concentration exceed the MCL at this well?", a coarse-grid simulation may report no even when the true concentration exceeds the threshold.

When is numerical dispersion acceptable?

For screening-level analysis — preliminary site assessment, go / no-go evaluation, is-there-a-problem questions — numerical dispersion's effects are usually acceptable because you're looking for order-of-magnitude understanding, not precise concentrations. The conservative bias on plume extent is often helpful at this stage. Run a coarse simulation, get a fast answer, identify where to invest more effort.

For quantitative work — regulatory compliance demonstration, remediation design, legal defense — numerical dispersion must be controlled. The remedies in IGW-NET:

Remedies in IGW-NET — in order of effort

The diagnostic workflow — plume against particles

Because particle tracking has no numerical dispersion, pairing the two analyses is one of the most useful diagnostic techniques in IGW-NET:

12.4.4 Molecular diffusion — rarely the main process

Diffusion is parameterized in IGW-NET by an activation flag plus three directional diffusion coefficients (D_xx, D_yy, D_zz) in the same Aquifer Attributes dialog as dispersivity. Defaults are all zero and off — diffusion is not active in a new model.

For most IGW-NET transport problems, diffusion is the smallest of the five processes and you can leave it at zero without meaningful error. At typical groundwater velocities, dispersion dominates diffusion by two to four orders of magnitude. There are three situations where you should turn diffusion on:

• Very slow-flow systems where you're modeling transport through confining units or aquitards. When groundwater velocity is near zero, diffusion is the main mass-transport process.
• Dual-domain transport (§12.7) — mass exchange between the mobile and immobile domains is driven by diffusion. If you're using dual-domain in IGW-NET, diffusion needs to be active.
• Long-timescale legacy problems — century-scale simulations where even small diffusion contributions can matter.

Typical values when you do need them: 10^-5 to 10^-4 m²/day in saturated porous media, with temperature correction possible through SEAWAT's viscosity package (§12.8). The default isotropic assumption (D_xx = D_yy = D_zz) is adequate for most purposes.

12.4.5 Sorption and retardation — two parameterization paths

Sorption represents reversible mass-exchange between the dissolved phase and the aquifer solids, slowing plume migration relative to groundwater flow. In IGW-NET, you activate sorption through the Sorption/Decay section of the Aquifer Attributes dialog. The platform offers two ways to specify the same physics — pick whichever matches your available data:

The visual consequence of turning sorption on: the plume moves slower than groundwater velocity would predict. A retardation factor of 5, for example, means the plume's center of mass advances at one-fifth the groundwater velocity. For a pulse release, the plume moves slower and also tails longer (sorbed mass slowly desorbs, maintaining low concentrations after the main body passes). Watch the animated simulation with and without sorption — the difference is immediately visible.

Two IGW-NET-specific choices worth being deliberate about:

• Soil bulk density default (265,000 g/m³). This is the platform's canonical value for saturated granular aquifer material. For rock aquifers or unusually dense/loose material, adjust: sandstone around 250,000, limestone around 270,000, loose alluvium around 180,000. An error here of 20% produces a corresponding 20% error in computed R from K_d.
• Species-specific sorption in multi-species simulations. Each species in a multi-species simulation (§12.5) has its own K_d — PCE sorbs strongly, VC weakly, benzene moderately. A single global K_d for all species is wrong; the Multiple Species dialog has per-species parameters for this reason.

12.4.6 First-order decay — half-life or λ

Decay represents irreversible mass loss — biodegradation, radioactive decay, hydrolysis, or any process that destroys mass rather than just moving it around. In IGW-NET, decay is activated in the same Sorption/Decay section as sorption, with two alternative parameterizations available (just like sorption):

The two parameterizations are mathematically equivalent: half-life = 0.693 / λ. IGW-NET computes the other when you supply either. Typical half-lives to use:

12.4.7 2D vs 3D — the sublayer question

Before going further with transport setup, one architectural decision deserves attention: do you need vertical resolution in the plume, or is a plan-view 2D simulation sufficient for your purpose? Both are legitimate; the right choice depends on what you're using the model for.

When 2D transport is appropriate

A plan-view 2D transport simulation — one conceptual layer, one sublayer, no vertical dispersion or diffusion — is a reasonable starting point and often the right final choice. It is:

• Fast and cheap — simulations run in seconds, so you can iterate through source configurations, dispersivity values, and scenarios quickly.
• Slightly conservative on horizontal extent — with no vertical dispersion to carry mass away from the plume axis, 2D plumes tend to be somewhat longer and more concentrated than equivalent 3D plumes. For screening purposes this is often the right direction to err.
• Easier to interpret visually — a single plume in plan view is the clearest way to convey "contamination goes here" to stakeholders.
• Sufficient for regional plumes that fill the full aquifer depth — if a long-standing nonpoint source has distributed a contaminant across the entire saturated thickness, 2D is physically appropriate.

Many legitimate modeling questions are answered in 2D: which direction does a plume go, roughly how far does it travel, does it reach a particular receptor, what's the order-of-magnitude concentration at that receptor. For site screening, regulatory scoping, or general conceptual understanding, 2D is not a compromise — it's often the right level of detail.

When 3D becomes important

The case for 3D modeling is strongest when you need the simulation to match specific measured concentrations, rather than just describing plume location and rough magnitude. Field measurements are inherently three-dimensional — monitoring wells sample specific screen depths, and concentration typically varies substantially with depth in a real plume. A contaminant released at the water table concentrates in the upper aquifer and gradually disperses downward; a well screened deeper in the aquifer will see lower concentrations than a shallow well at the same horizontal location.

When you compare a 2D simulation to these depth-specific measurements, you face a structural problem: the 2D model has averaged across depth, so its simulated concentration lies somewhere between the shallow-well value and the deep-well value. Your simulated numbers don't match either observation, and no amount of parameter tuning — K, dispersivity, sorption — can fix this. You're comparing a depth-averaged simulation to depth-specific data, and the mismatch is inherent to the 2D formulation. 3D is what lets the model produce depth-resolved concentrations that can actually match what the monitoring wells measured.

Concrete indicators that 2D is reaching its limit:

• You have monitoring data from multiple wells at different screen depths at the same horizontal location, and they show different concentrations
• You're doing transport calibration and the fit to measurements is systematically wrong in a way that doesn't respond to parameter changes
• Your problem involves DNAPL or dense-plume sinking, where the vertical descent is a central part of the physics
• You have a partially-penetrating pumping well whose capture zone depends on which depth interval is screened
• You need to differentiate risk to shallow-aquifer receptors (domestic wells, wetlands) from deep-aquifer receptors (public-supply wells)

The IGW-NET transition — sublayers with water-table-as-top

Chapter 10 introduced the distinction between conceptual layers (real geologic units) and sublayers (numerical subdivision for vertical resolution). For transport, sublayers are the common mechanism for moving from 2D to 3D within a single aquifer — you don't necessarily need multiple conceptual layers. Adding 5-10 sublayers to a single conceptual layer gives you 3D transport resolution within that aquifer.

IGW-NET has a particularly efficient way to handle this, controlled by flags in Simulation Settings. The mechanism is covered in detail in Ch. 7 §7.4.3 — this section summarizes it for the transport context:

How many sublayers for transport?

12.5 Multi-Species Simulation in IGW-NET

Most plumes in the real world involve more than one dissolved species — a primary contaminant plus its degradation daughters, multiple co-contaminants from the same source, an electron donor introduced for remediation alongside the target contaminant. IGW-NET's Multiple Species dialog is where you tell the platform which species you want simulated and how they interact. This section walks through the dialog itself.

12.5.1 When to leave single-species — a practical checklist

Stay in single-species mode (the default) when:

• You only care about the parent compound and daughters are irrelevant to your decision — like the Mancelona case (§12.9.5), where TCE alone was the regulatory concern.
• Your contaminant is known to be conservative or degrades slowly enough that daughters are negligible over your simulation period (PFAS, many inorganic contaminants).
• You are doing an initial screening simulation where additional species would slow things down without changing the main question.

Engage multi-species mode when:

• Daughters are independently regulated. Vinyl chloride's MCL (2 ppb) is more stringent than PCE's (5 ppb), so a PCE plume that dechlorinates to VC can present greater risk than the original release suggests. Tracking each daughter separately is the only defensible approach for regulatory work on chlorinated solvents.
• Electron acceptors are limiting. For BTEX under natural attenuation, oxygen or nitrate may run out before hydrocarbons do — the plume's fate depends on which acceptor is present where. Simulating hydrocarbon alone with first-order decay misses this entirely.
• You're designing an ED/EA remediation. Injecting ethanol to reduce Cr(VI) requires tracking both species: where they overlap, reaction happens; where they don't, the Cr(VI) persists. Single-species cannot capture this.

12.5.2 Opening the Multiple Species dialog — three choices

From Domain Attributes → Sorption/Decay, click Multiple Species. The dialog opens with three radio-button choices at the top that set the overall approach. These three choices dictate everything else in the dialog:

12.5.3 Per-species parameter blocks

Regardless of which of the three approaches you chose, every species in the simulation gets its own parameter block inside the dialog — because sorption, decay, and dispersivity differ by species. PCE sorbs substantially; VC barely sorbs. PCE in anaerobic conditions has a half-life of years; VC under the same conditions has a half-life of months. Assuming one global set of parameters for all species is one of the most common and damaging mistakes in multi-species modeling.

The per-species block mirrors the single-species Aquifer Attributes fields from §12.4: the same flags (sorption active, decay active) and the same alternative parameterizations (K_d vs R, half-life vs λ). You fill them per species, not once globally. Predefined modules populate reasonable defaults; custom approaches leave them blank for you to fill.

12.5.4 Sources in multi-species mode

Source configuration (§12.3) extends naturally to multi-species. Every source location — a source zone, an injection well, a polluted stream — has a concentration field per active species. A zone can release pure PCE (PCE concentration = source value, TCE/DCE/VC concentrations = 0), or a mixed plume (PCE, TCE, and DCE all non-zero to represent an already-degraded release), as your conceptual model requires.

The rule from §12.2.3 still applies per species: an external source contributes that species' concentration only when water is flowing into the aquifer at that segment, and only if that species has a non-zero concentration at that boundary.

12.5.5 How reactions combine with the physical processes

Inside the solver, reactions and the five physical processes from §12.4 operate simultaneously. At every cell at every time step, the solver: advects each species with flow, applies dispersion and diffusion per species, applies per-species sorption, applies per-species decay (which for a parent species may transfer mass to a daughter), and applies any inter-species reactions from the module. The order matters for numerics but not for the user — IGW-NET handles the sequencing automatically. What you control: which species, how they sorb, how they decay, how they react. What the platform handles: putting those definitions into the advection-dispersion-reaction solution at every time step.

12.6 The Five Predefined Reaction Modules

IGW-NET's Multiple Species dialog includes five pre-built reaction modules — each one is a species list plus the reaction topology connecting them, packaged so that you select the module and immediately have the right species and right linkages. You still supply the rate constants, source concentrations, and per-species sorption, but the wiring is done. This section covers what each module includes, when to use it, and the specific inputs it expects.

12.6.1 Module 1 — BTEX Instant Aerobic Decay

What you get when you select it: A species list of five species — benzene, toluene, ethylbenzene, xylene, and oxygen — with an instantaneous reaction linking each hydrocarbon's consumption to oxygen depletion at the stoichiometric mass ratio. No kinetic rates to supply; the reaction proceeds as fast as mass allows.

What you fill in: Source concentrations for each hydrocarbon species present at your source location; ambient oxygen concentration (typically the dissolved-oxygen saturation ~8-10 mg/L in shallow freshwater systems, lower in deeper anaerobic zones). Each hydrocarbon gets its own sorption/decay parameters — you enter these per species in the dialog's parameter block.

When it's the right module: Your problem is a fresh BTEX plume in an aerobic aquifer, you want a screening-level simulation, and the timescale of interest is long enough that you can treat hydrocarbon-oxygen reaction as instantaneous. If the site has O₂ > 2 mg/L in the background groundwater, this is your starting module. Switch to Module 2 when you see electron-acceptor sequencing (a site with a well-developed anaerobic core).

12.6.2 Module 2 — BTEX Kinetic Decay with Multiple Electron Acceptors

What you get when you select it: A species list that includes the hydrocarbons plus five electron acceptors — O₂, NO₃⁻, Fe³⁺, SO₄²⁻, and CO₂ (for methanogenesis) — with kinetic rate equations for degradation via each pathway. The pathways are sequenced: as each acceptor is consumed, the next takes over.

What you fill in: For each hydrocarbon, a first-order rate constant for each of the five pathways (five rates per hydrocarbon — ten total for benzene-plus-toluene, etc.). Ambient concentrations of each electron acceptor. Source concentrations per hydrocarbon.

The payoff — why this module is worth the extra configuration: This is the module that reproduces the classic BTEX-plume zonation that regulators and site assessors expect to see. When you run it, the simulation shows an aerobic fringe around the plume edges where oxygen is depleted first; an iron-reducing and sulfate-reducing core where those acceptors dominate; and potentially a methanogenic center for very old plumes. Single-pathway models (including Module 1) cannot reproduce this zonation. When your field data show different redox signatures at different distances from the source, Module 2 is how you match it.

12.6.3 Module 6 — PCE Sequential Decay

What you get when you select it: Five species — PCE, TCE, DCE, VC, ethylene — with first-order parent-child links from PCE down through the dechlorination chain to ethylene. The topology is pre-wired: PCE decays to TCE at rate k₁, TCE to DCE at k₂, DCE to VC at k₃, VC to ethylene at k₄.

What you fill in: Four rate constants (k₁ through k₄), source concentrations per species, and per-species sorption parameters. A typical setup has PCE entering the source (k₁ fastest) and the downstream intermediates lower concentration, so source PCE is usually the only non-zero source species.

When it's the right module: Your site is a chlorinated-solvent release (dry cleaner, machining facility, military installation), conditions are or can be made anaerobic, and you need to track each daughter separately for regulatory purposes. Remember DCE and VC have lower MCLs than PCE (70 and 2 ppb vs 5 ppb respectively), so "the PCE is decaying" is not automatically good news — it may be producing more-regulated compounds. This module is the standard representation for such sites.

12.6.4 Module 7 — PCE/TCE Aerobic plus Anaerobic

What you get when you select it: Module 6's species list plus oxygen, with both the anaerobic dechlorination pathway (Module 6) and aerobic degradation pathways for daughter products operating simultaneously. TCE and the subsequent daughters can degrade aerobically when oxygen is present, independent of the anaerobic chain.

What you fill in: Module 6's four anaerobic rates, plus aerobic rate constants for each daughter species that has aerobic degradation active, plus ambient and source oxygen concentration. Sorption and source concentrations per species as before.

When it's the right module: Your plume spans both aerobic and anaerobic zones — a common situation when a chlorinated-solvent plume has a core anaerobic area (near source) transitioning to aerobic conditions downgradient. This is the module that handles both regimes in one simulation, which matters for mass balance and for defending the model to regulators who will ask "which pathway is removing mass here?"

12.6.5 Single-Pair Instant EA/ED

What you get when you select it: Two species — one designated as electron acceptor (EA), one as electron donor (ED) — with an instantaneous reaction between them when they coexist in a cell. The reaction consumes both at a stoichiometric mass ratio you specify.

What you fill in: Which species is EA and which is ED; the stoichiometric mass ratio (how many mg of donor are consumed per mg of acceptor); source concentrations for each; per-species sorption.

The canonical use case: In-situ Cr(VI) reduction by injected ethanol. You set Cr(VI) as the EA, ethanol as the ED, and an appropriate stoichiometric ratio. Inject ethanol at a well (set the well as an injection well with ethanol concentration); let the simulation run; observe where the ethanol plume overlaps the Cr(VI) plume and watch both get consumed in the overlap zone. This is how you size injection wells, plan injection schedules, and defend the design of an in-situ remediation scheme. Other EA/ED pairs — nitrate reduction by acetate, perchlorate reduction by lactate — use the same module with different inputs.

12.7 Dual-Domain Transport in IGW-NET

For fractured rock, macropores, karst systems, and strongly heterogeneous aquifers, single-domain transport (what §12.4 covers) gives plumes that move too fast and don't tail right. IGW-NET's dual-domain option splits the medium into a mobile (flowing) and an immobile (stagnant) domain with diffusive exchange between them. This section walks through when to use it, how to activate it in IGW-NET, and the three parameters you need to supply.

12.7.1 The symptom — long tailing that single-domain cannot match

The practical indicator that you need dual-domain in IGW-NET: you have a calibrated flow model, you're running single-domain transport, and the simulated breakthrough curve at your downgradient monitoring well has a fundamentally wrong shape. Specifically: the simulation shows a clean rise-to-peak-to-decline pattern, but your actual monitoring data show a steep rise, then concentrations that stay elevated or decline much more slowly than the simulation predicts. That long persistent tail is the signature of matrix back-diffusion, and no single-domain parameterization will reproduce it — increasing dispersivity, tweaking sorption, or adjusting decay all fail.

When you see this symptom in your IGW-NET simulation, dual-domain is what's missing.

12.7.2 Activating dual-domain — the Advanced Transport package

Dual-domain is not in the basic Aquifer Attributes dialog — it lives in the MT3DMS Advanced Transport package settings within IGW-NET's Simulation Settings. To activate it:

12.7.3 The three parameters and typical values

12.7.4 When to reach for dual-domain in IGW-NET

12.8 Variable-Density Flow with SEAWAT

For problems where dissolved concentration or temperature significantly changes water density, the coupling between flow and transport runs in both directions: the plume depends on the flow field, and the flow field depends on the plume's density. IGW-NET integrates the SEAWAT engine to handle this coupling. This section covers when you need it, how to switch IGW-NET to SEAWAT mode, and what the setup requires.

12.8.1 When variable density actually matters

For the vast majority of groundwater problems, water density is effectively constant and standard MODFLOW-plus-MT3DMS handles everything. You need SEAWAT when:

• Total dissolved solids exceed roughly 1,000 mg/L. Below that, density variation is negligible and you're wasting computation in SEAWAT.
• The interface between fresh and dense fluids matters. Seawater intrusion in coastal aquifers is the flagship case — the 1.02-ish specific gravity of seawater drives a dense wedge under fresh groundwater, and the interface's position is what regulators and water utilities want to know.
• Brine migration is the question. Injection of brine for disposal, natural brine from evaporite dissolution, or historical brine plumes from oil-field operations.
• Thermal plumes couple to flow. Industrial cooling-water discharge, geothermal installations, aquifer thermal energy storage (ATES) — all involve temperature differences large enough to affect density.

If your problem matches none of these, keep the standard engine. SEAWAT is a solver choice, not a default — you activate it deliberately when density matters.

12.8.2 Switching IGW-NET to SEAWAT mode

12.8.3 What you see — the density-dependent plume

The visual payoff of SEAWAT mode in IGW-NET: you watch the plume and the flow field evolve together, each influencing the other. In a coastal seawater-intrusion simulation, the animation shows the saltwater wedge advancing inland and deepening as sea-level or pumping pressures change; the flow field above the interface is freshwater-dominated and flows seaward; the flow field below the interface is saltwater-dominated and moves differently. The interface is typically sharp enough that you can trace it in cross-section.

For a brine plume, the animation shows the plume sinking as it moves — heavier fluid dives beneath the overlying freshwater, producing a layered structure that single-density transport would miss. For thermal plumes, warm-water plumes from surface sources tend to float; cold-water plumes from deeper injections tend to sink. SEAWAT makes all of these visible in a way that's convincing to stakeholders who need to understand the physics.

12.9 Reading Transport Results

Transport simulations produce richer output than flow alone. In addition to head contours and velocity vectors, you now see plumes in plan view and cross-section, breakthrough curves at monitoring points, and a transport-specific mass balance. Interpreting all of this is where transport modeling becomes useful.

12.9.1 Plume plan views

The plume plan view is the most immediate output of transport simulation. What to look for:

• Plume trajectory — does the plume go where flow predicts? Advection should dominate the plume's axis; any deflection usually indicates heterogeneity or source positioning.
• Plume shape — elongated in the flow direction (characteristic of anisotropic dispersion with αL > αT); rounded (isotropic or strong transverse dispersion); complex (heterogeneous aquifer, multiple sources, or boundary interaction).
• Contour spacing — tight contours at the leading edge (sharp front, advection-dominated); broad contours (dispersion-dominated or dual-domain release).
• Multiple plumes from multiple sources — overlapping where sources are close; separated where far apart; merging downgradient into a broader plume.

12.9.2 Cross-sections — vertical structure of the plume

For layered systems and for any plume with vertical structure, cross-sections are essential. Key questions to ask:

• Is the plume staying in the upper aquifer, or migrating downward through confining units?
• If a DNAPL, is the simulation showing the expected vertical descent?
• Are pumping wells capturing the plume, or drawing it deeper?
• For dual-domain or matrix diffusion, is the plume distributing vertically as expected?

12.9.3 Breakthrough curves at monitoring wells

Chapter 9 §9.8.3 introduced breakthrough curves — concentration at a monitoring well plotted against time. They answer quantitative questions that plan views cannot:

• Arrival time — when does concentration first rise above background?
• Time to MCL exceedance — when does concentration exceed the regulatory maximum contaminant level?
• Peak concentration and timing — what is the highest concentration the well sees, and when?
• Decline rate — how fast does concentration fall after the peak?
• Tailing — how long do low-level concentrations persist? Strong tailing suggests dual-domain behavior or slow desorption.

Breakthrough curves are what regulators most often ask for — they are the quantitative summary of risk at a receptor.

12.9.4 Transport mass balance

The water balance from Ch. 8 §8.4 has a transport analog: the mass balance tracks contamination entering, leaving, and accumulating in the domain. Checking it is essential for any serious transport calibration.

Mass inflow categories:

• Internal source release (from source zones)
• Injection well mass input
• Polluted recharge mass input
• Mass inflow through boundaries

Mass outflow categories:

• Mass extracted via pumping wells
• Mass discharged at one-way drains and losing streams
• Mass outflow through boundaries
• First-order decay (mass destroyed, not transported)
• Reaction network consumption (for multi-species simulations)

Storage change:

• Mass in the dissolved phase (mobile plus immobile for dual-domain)
• Mass in the sorbed phase

In a well-constructed steady-state transport simulation, the mass balance closes: inflows = outflows + storage change. Imbalances larger than 1-2% usually indicate convergence issues or configuration problems worth investigating.

12.9.5 A real example — the Mancelona TCE plume

The Mancelona case (northern Michigan) is a real TCE contamination investigation. A historical source area near a former Wickes Manufacturing facility released TCE that migrated for decades through a glacial-outwash aquifer toward local water-supply wells. The case study models TCE as a single species — daughter products were not part of the investigation objectives, so the simulation tracks TCE alone rather than engaging a multi-species reaction module. The workflow is a good example of a focused single-species transport model: regional base model, submodel for the source area with refined grid, single-species TCE transport with sorption and first-order decay, forward particle tracking from the source area to confirm the flow trajectory, and calibration against observed plume concentrations. See the case study documentation for the complete workflow, and §12.6 of this chapter for when multi-species modules are worth engaging.