Transport Processes

What are the four things that happen to a contaminant once it enters groundwater?

What you’re watching The same initial plug of contaminant evolving under each process in turn: under advection it travels downstream as a compact slug; under pore-scale dispersion it smears and blurs; under sorption it lags behind — sitting upstream of where advection alone would carry it; under first-order decay it moves but fades as mass is lost.

The mechanism Advection carries solute with the flowing water. Dispersion spreads it through velocity variation and mixing, lowering the peak concentration. Sorption partitions solute onto the solid grains, retarding the plume so it moves slower than the water (retardation factor R). Decay — radioactive, chemical, or microbial — destroys mass over time.

Why it matters Every cleanup forecast and risk assessment is some combination of these four. Whether a plume reaches a well at a dangerous concentration depends on how fast (advection), how spread (dispersion), how delayed (sorption), and how much survives (decay).

IGW-NET Run the same release with each process acting alone and the abstract terms of the transport equation become four visibly different plumes. IGW-NET lets you ‘experience’ dispersion and retardation directly — simulating each mechanism is the act of seeing it.

Where does most ‘dispersion’ really come from at the field scale?

What you’re watching A plume in a heterogeneous aquifer spreading far more than pore-scale dispersion alone could explain — the conductivity variations, not molecular mixing, are doing the spreading.

The mechanism Pore-scale dispersion is tiny at field scale. The dominant spreading comes from velocity contrasts between high- and low-K zones that pull the plume apart — ‘macrodispersion.’ What looks like dispersion is mostly unresolved advection through heterogeneity.

Why it matters Using a small lab-measured dispersivity drastically underpredicts real plume spread — a core reason early cleanup designs fell short.

IGW-NET Watch a plume disperse through a heterogeneous field, then through a uniform one; at a glance you see how much of ‘dispersion’ is really heterogeneity.

Does heterogeneity always behave like simple dispersion?

What you’re watching A heterogeneous-aquifer plume whose fingered, irregular shape doesn’t match the smooth bell shape a single dispersion coefficient would produce.

The mechanism True heterogeneous transport produces fingering, channeling, and tailing — non-Fickian behavior. A single dispersion coefficient assumes smooth Gaussian spreading, which the real plume violates, especially early and in strongly heterogeneous media.

Why it matters Treating heterogeneity as plain dispersion hides the fast fingers that deliver contamination early — the ones that matter for a downstream well.

IGW-NET Comparing the fingered real plume with a smooth dispersion model side by side makes the limits of the effective-dispersion idea visible.

Can we replace a messy heterogeneous aquifer with a simple ‘effective’ one?

What you’re watching Side by side: the real heterogeneous aquifer (left) where the plume fingers and channels through the conductivity field, and the macrodispersion model (right) that replaces it with a uniform aquifer and a large effective dispersivity — a smooth, spread blob.

The mechanism The macrodispersion model captures the overall spreading (the plume’s second moment) by lumping heterogeneity into one large effective dispersivity. But it smooths away the fingering and overstates dilution — the blob has roughly the right extent but the wrong internal structure and peak concentration.

Why it matters Effective models are convenient and widely used, but relying on them can underestimate peak concentrations and miss the fast fingers — exactly where risk lives. Knowing when the effective model is and isn’t valid is a central research question.

IGW-NET Running the true heterogeneous transport and its macrodispersion equivalent at the same time shows precisely what the effective model keeps and what it discards — the kind of model-validity experiment a real-time digital laboratory makes routine.