🔬 physics

The Statistical physics of unsaturated soil water: kinetic theory and non commutative pore water dynamics

This paper presents a multiscale statistical-mechanical framework that unifies the dynamics of unsaturated soil water into a single kinetic continuum equation, demonstrating that classical models like Richards' equation and phenomena such as preferential flow and hysteresis emerge as specific limits or geometric consequences of a pore-resolved Damkohler number governing non-equilibrium pore occupancy.

Original authors: Riccardo Rigon

Published 2026-07-13

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Riccardo Rigon

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine soil not as a muddy lump, but as a giant, 3D maze made of billions of tiny, hollow tubes of different widths. Some are as thin as a human hair; others are as wide as a drinking straw. Inside this maze, water is trying to find its way.

For over a century, scientists have tried to describe how this water moves using a single, famous rule called the Richards equation. Think of this equation as a "slow-motion" rulebook. It assumes that if you pour water on the soil, the water instantly finds the most comfortable, energy-saving spot for every single drop. It assumes the water is always perfectly happy, always in equilibrium, and that the only thing that matters is how much water is in the soil.

The Big Discovery: The Soil Has a Memory (and a Speed Limit)

This paper, by Riccardo Rigon, argues that the old rulebook is incomplete. It's like trying to describe a chaotic mosh pit by only looking at the average crowd density. You miss the fact that people are running, bumping into each other, and getting stuck in specific spots.

The paper introduces a new, more detailed view. Instead of just asking "How wet is the soil?", it asks: "Which specific tubes are full, and which are empty?"

The authors propose a new "kinetic" theory. They treat the water in the soil like a crowd of people in a building with many rooms.

The Microscale (The Hallway): Water moves between tubes based on physics (Hagen–Poiseuille rates). It's like water rushing through a pipe.
The Mesoscale (The Room): When you look at a whole room (a tiny chunk of soil), the water is constantly swapping places, trying to settle into the most comfortable arrangement.
The Macroscale (The Building): When you zoom out to the whole field, you see the water flowing from one room to the next.

The "Damköhler Number": The Soil's Speedometer

The most exciting part of this paper is a single, magic number they call the Damköhler number (Da). Think of this as a "Speedometer" for the soil. It compares two speeds:

How fast the water wants to rearrange itself to find the perfect spot (relaxation).
How fast the rain is hitting the ground (forcing).

If the rain is slow (Da is small): The water has plenty of time to rearrange itself. It finds the perfect spots, just like the old Richards equation predicted. The soil behaves like a calm, slow-moving fluid.
If the rain is a sudden downpour (Da is large): The water gets dumped in faster than it can rearrange. It doesn't have time to find the tiny, comfortable tubes. Instead, it gets stuck in the big, fast tubes, creating "preferential flow" (like water shooting through a firehose).

The "Preferential Flow" Mystery Solved?

For a long time, scientists thought "Richards flow" (slow, steady) and "Preferential flow" (fast, chaotic) were two totally different things that needed two different rulebooks.

This paper argues NO. They are the same equation, just behaving differently depending on the speed of the rain.

Richards flow is what happens when the soil has time to catch its breath.
Preferential flow is what happens when the soil is overwhelmed.

It's not that the soil changes its mind; it's that the speed of the event changes the outcome. The paper suggests that if you know the speed of the rain and the size of the tubes, you can predict exactly when the soil will switch from "calm" to "chaotic."

The "Hysteresis" Loop: A Geometric Ghost

You've probably noticed that soil holds water differently when it's drying out compared to when it's getting wet. This is called hysteresis.

Old Idea: Scientists thought this was because the soil had "bistable" tubes—like light switches that get stuck in the "on" or "off" position depending on history.
New Idea: The paper argues this isn't about stuck switches. It's about geometry and time.

Imagine walking a path in a forest. If you walk slowly, you can step over every twig. If you run, you might trip over a branch and take a different path. If you walk in a circle (wet then dry), you might not end up exactly where you started because the "running" path was different from the "walking" path.

The authors call this a geometric phase (or holonomy). It's like a "geometric ghost." The soil remembers the path it took to get to its current wetness, not just the wetness itself. The paper suggests that the "size" of this memory (the hysteresis loop) grows with the square of the rain intensity. If you double the rain speed, the memory effect gets four times bigger. This is a specific, testable prediction that the old "stuck switch" models cannot explain.

What is Proven, What is Suggested, and What is Ruled Out

Ruled Out: The paper explicitly argues against the idea that "preferential flow" and "Richards flow" are two separate, unrelated phenomena. They are one continuous crossover. It also rules out the idea that hysteresis is caused by individual pores getting "stuck" in a binary state; instead, it's a result of the whole network's geometry and the speed of the forcing.
Proven (in the paper's math): The authors have mathematically proved that their new equation preserves the laws of physics (like mass conservation) and that the system always moves toward a lower energy state (the H-theorem). They have shown that the new equation can mathematically turn into the old Richards equation if you slow everything down enough.
Suggested/Simulated: The specific prediction that the hysteresis loop area grows with the square of the rain intensity ( $H \propto I^2$ ) is a conjecture based on their theory. They suggest this could be tested in a lab. They also suggest that the "field capacity" (the point where soil stops draining) is actually a "percolation threshold" (a point where the water network breaks apart), which is a plausible idea supported by their model but needs more real-world verification.
Simulated Only: The idea that the "non-commutativity" (the fact that wetting then drying gives a different result than drying then wetting) is the root cause of hysteresis is demonstrated in their math and in computer simulations of a small network of tubes.

The Bottom Line

This paper doesn't just offer a new formula; it offers a new way of seeing the world. It suggests that the soil isn't a static sponge, but a dynamic, time-sensitive maze. Whether the water flows slowly and evenly or rushes through in chaotic streams depends entirely on the speed of the rain compared to the speed of the soil's internal rearrangement.

The authors are confident that their new equation is the "master key" that unlocks both the slow, steady flow and the fast, chaotic flow, unifying them into a single, beautiful, and mathematically rigorous story. They invite scientists to test their prediction: if you measure the hysteresis loop at different rain speeds, it should grow quadratically. If it does, the "geometric ghost" is real. If not, the theory needs a rewrite.

Technical Summary: Statistical Physics of Unsaturated Soil Water

Problem Statement
The dynamics of unsaturated porous media are inherently irreversible, characterized by non-equilibrium configurations that depend on the history of forcing (wetting vs. drying). Classical models, primarily Richards' equation, assume local thermodynamic equilibrium and a single scalar state variable (water content $\theta$ or matric potential $\psi$ ). This framework fails to explain:

Hysteresis: Identical macroscopic states corresponding to different internal configurations depending on the wetting/drying path.
Preferential Flow: The coexistence of slow, capillary-controlled redistribution and rapid, bypass flow through large pores during intense forcing, which violates the single-domain assumptions of Richards' equation.
Rate Dependence: The inability of rate-independent models to capture the dependence of hysteresis and flow regimes on the speed of forcing.

Existing dual-domain or dual-permeability models address these issues phenomenologically by introducing additional state variables and exchange terms without deriving them from a common physical description.

Methodology
The paper develops a statistical-mechanical theory by passing through three scales, treating the pore-occupancy fraction $g(r, x, t)$ (the fraction of pores of radius $r$ filled with water at position $x$ and time $t$ ) as the fundamental state variable.

Microscale (Kinetic Rates):
- Inter-pore water transfers are governed by Hagen–Poiseuille rates $\kappa(r)$ and a driving potential $\Phi(r, r')$ , defined as the difference in chemical potentials between pore classes (incorporating capillary and gravitational terms).
- A state-dependent rate constant $\kappa(r, g)$ is derived, transitioning from Lucas–Washburn dynamics (fast initial invasion) to Hagen–Poiseuille flow (steady state) as the filling fraction $g$ increases.
Mesoscale (Master Equation):
- Averaging over a Representative Elementary Volume (REV) yields a master equation for $g$ .
- The equation includes gain ( $G$ ) and loss ( $L$ ) terms representing water exchange between pore classes. These terms utilize a "Stosszahlansatz" (molecular chaos hypothesis), assuming statistical independence of donor and receiver occupancies, leading to a bilinear form $g(r')[1-g(r)]$ .
- The system is shown to be an Onsager gradient flow, descending a Gibbs free energy functional $F[g]$ with a state-dependent mobility. This ensures the H-theorem ( $dF/dt \leq 0$ ) for isothermal, unforced systems.
Continuum Scale (Field Equation):
- Contracting the REV to a point promotes $g$ to a continuum field.
- The resulting fundamental equation is a balance law:
  $\partial_t g + \nabla \cdot \mathbf{F} = C[g] - E - T$
  where $C[g]$ is the intra-REV redistribution operator, $\nabla \cdot \mathbf{F}$ is the transport divergence (inter-REV flux), and $E, T$ are sinks (evaporation, root uptake).
- The flux $\mathbf{F}$ is non-local in pore radius, coupling different pore classes via the connectivity matrix $C(r, r')$ .

Key Contributions and Results

Unification of Flow Regimes: The theory introduces a single dimensionless control parameter, the pore-resolved Damköhler number $Da(r, x) = \tau_{redis}(r) / \tau_{forcing}(x)$ , which compares the internal relaxation time to the forcing time scale.
- $Da \ll 1$ (Quasi-static): The system tracks local equilibrium. The Chapman–Enskog reduction of the kinetic equation recovers Richards' equation, with matric potential $\psi$ and hydraulic conductivity $K(\theta)$ emerging as limits.
- $Da \gtrsim 1$ (Kinetic): Redistribution cannot keep pace with forcing. Water occupies geometrically accessible pathways (large pores) rather than thermodynamically optimal ones, naturally reproducing preferential flow without ad hoc dual-domain assumptions.
Geometric Origin of Hysteresis: Hysteresis is not attributed to per-pore bistability (e.g., ink-bottle effects) but to the holonomy of a "forcing bundle."
- The wetting ( $L_W$ ) and drying ( $L_D$ ) operators do not commute ( $[L_W, L_D] \neq 0$ ) due to the asymmetry between non-selective filling (precipitation) and selective emptying (drainage/evaporation constrained by topology).
- This non-commutativity creates a geometric phase (residual state) after a closed forcing cycle.
- Prediction: The hysteresis loop area $H$ scales quadratically with forcing intensity ( $H \propto I^2$ ) for slow cycles, a signature absent in rate-independent models.
Constitutive Laws from Microstructure:
- Hydraulic conductivity $K(\theta)$ is derived as a transport coefficient (spectral limit of the mobility kernel) rather than a fitted parameter.
- The theory recovers classical limits: Capillary-bundle models (diagonal limit), critical-path models (spectral limit near percolation), and dynamic capillary-pressure models (one-moment approximation).
- Field capacity is identified as the percolation threshold where the water network fragments ( $\theta_c$ ), causing macroscopic conductivity to vanish.
Irreducible State Variable: Outside the quasi-static limit, the scalar water content $\theta$ is insufficient to characterize the system. The full distribution $g(r, x, t)$ is the irreducible state variable.

Significance and Claims
The paper claims to provide a constructive, thermodynamically grounded framework that unifies previously distinct phenomenologies (Richards' flow, preferential flow, hysteresis) into a single kinetic equation.

Theoretical Rigor: The theory is presented as a porous-media analogue of the Boltzmann equation, with $g(r)$ corresponding to the velocity distribution $f(v)$ , and the Chapman–Enskog reduction mirroring the derivation of Navier–Stokes equations.
Predictive Power: All inputs (pore-size distribution $f(r)$ , connectivity $C(r, r')$ , tortuosity $\bar{\tau}$ ) are geometric properties measurable from micro-CT images, requiring no calibration against macroscopic retention or conductivity data.
Falsifiability: The theory makes specific, testable predictions, most notably the $H \propto I^2$ scaling of hysteresis loop area with forcing intensity and the existence of a derived threshold intensity for the onset of preferential flow based on soil structure and fluid properties.
Scope: The framework is generalizable to other systems with constrained, history-dependent transitions (e.g., plastic deformation, magnetic hysteresis) and can be extended to include adsorptive, osmotic, or thermal terms by refining the driving potential $\Phi$ without altering the core operator algebra.

The paper concludes that the dichotomy between equilibrium and non-equilibrium flow is not a matter of choosing different models, but of recognizing the continuous crossover controlled by the Damköhler number within a single, unified kinetic theory.

Technical Summary: Statistical Physics of Unsaturated Soil Water

More like this