A new study has revealed that tiny areas where rough rock fracture surfaces touch can dramatically alter how pollutants and other dissolved substances move underground. The finding challenges the common assumption that solute transport through a fracture can be predicted mainly from flow velocity and average aperture—the typical gap between two rock surfaces. Instead, the research shows that the geometry of the fracture, particularly the percentage of surfaces in direct contact, may determine whether a dissolved chemical moves rapidly through the subsurface or becomes trapped for long periods. The results are important for groundwater protection, contaminant cleanup, geothermal engineering, and the long-term safety assessment of underground nuclear-waste repositories. Published in Hydrogeology Journal, the study used high-resolution three-dimensional numerical simulations to examine how rough fracture surfaces guide water and dissolved solutes through complex microscopic pathways.
Natural rock fractures are not smooth, uniform channels. Their walls contain ridges, depressions, constrictions, and irregular openings created by geological stress and mechanical weathering. When two rough surfaces meet, some regions may be tightly pressed together while others remain open enough for water to pass. This produces a highly uneven flow field. Fast-moving streams can form through relatively wide channels, while stagnant or slowly circulating zones develop behind protrusions, beside contact points, or within recirculation regions. A dissolved substance entering such a fracture therefore does not experience a single representative velocity. Instead, different portions of the solute cloud travel at different speeds, and some particles may spend much longer in the fracture than others. That unequal distribution of travel times is a major cause of non-Fickian transport, a behavior characterized by early arrivals, broad breakthrough curves, and persistent late-time tailing.
To isolate the effect of surface contact, the researchers generated four three-dimensional fracture models based on self-affine roughness, a mathematical description commonly used to reproduce the scale-dependent irregularity of natural geological surfaces. The modeled contact areas ranged from zero to 17.9 percent. In a fracture with no direct contact, the opposing surfaces form a continuously connected void. As the contact area increases, the available flow space becomes more fragmented and tortuous. The simulations then tracked conservative solute movement under flow velocities ranging from 0.05 to 3 millimeters per second. These conditions allowed the team to compare slow and fast transport regimes while keeping the focus on the interaction between hydraulic forcing and fracture geometry. The resulting breakthrough curves recorded how the concentration at the outlet changed over time after solute entered the modeled fracture.
The simulations showed that increasing contact area strongly enhanced solute retention. At the highest modeled contact level, direct surface intersections narrowed the main flow routes and created larger regions in which water moved slowly or circulated locally. Solute entering these zones could later diffuse or be advected back into faster channels, producing a prolonged concentration tail after the main breakthrough event. This means that a fracture can deliver a contaminant to a downstream location relatively quickly while continuing to release smaller amounts for much longer. Such behavior complicates risk assessments because the first arrival and the total duration of contamination may be governed by different parts of the fracture network. A model that predicts only the average travel time could therefore underestimate the persistence of contamination even if it reproduces the peak concentration reasonably well.
The researchers quantified the degree of anomalous transport using the exponent β in a continuous-time random-walk framework. In this approach, particles are treated as making successive movements separated by waiting times that can vary widely. A truncated power-law distribution is used to represent the broad range of residence times while still allowing the system to transition toward more ordinary transport at sufficiently long times or under different conditions. Lower β values indicate stronger anomalous behavior and a greater influence of long waiting times. Across the contact-area scenarios, the reported β value declined from 1.79 in the fracture with no contact toward 0.56 at 17.9 percent contact. The latter represents strongly anomalous transport, in which a substantial portion of the solute remains influenced by slow pathways and retention zones.
Flow velocity produced the opposite trend. As the imposed velocity increased, transport became progressively less anomalous, with β values rising from approximately 0.2 to 1.2 across the tested velocity range. Faster flow can reduce the relative time available for solute to enter and remain in stagnant regions. It can also increase the dominance of advective transport, allowing particles to pass through the fracture before diffusion and recirculation substantially redistribute them. The finding does not mean that high velocity eliminates irregular transport altogether. Rather, it indicates that stronger through-flow can partially overwhelm the effects of geometric trapping. At low velocities, by contrast, the solute has more opportunity to exchange between fast channels and slow zones, magnifying the contrast in travel times and extending the breakthrough tail.
The study compared two ways of representing the simulated breakthrough curves: the classical advection–dispersion equation, or ADE, and the continuous-time random-walk model with a truncated power-law waiting-time distribution, known as CTRW-TPL. The ADE assumes that spreading can be described using averaged advection and dispersion terms, generally producing a relatively compact concentration distribution. This framework is powerful and computationally efficient, but it can struggle when transport is controlled by multiple flow speeds, dead-end regions, or long-lived recirculation. The CTRW-TPL model instead incorporates a distribution of particle waiting times, making it better suited to heavy-tailed breakthrough behavior. According to the study, the improvement in overall statistical fit was small: the CTRW-TPL model increased the coefficient of determination, R², by only about 0 to 0.008 compared with the ADE.
That modest numerical improvement is one of the study’s most important messages. A model can produce a similar global R² while representing the underlying physics very differently. If most measured concentrations are concentrated around the breakthrough peak, a statistical score may give relatively little weight to the low-concentration tail, even though that tail is crucial for predicting long-term contaminant persistence. The CTRW-TPL model’s advantage was therefore not simply that it fit every point dramatically better, but that it offered a physical mechanism for the heavy-tailed behavior observed in high-contact fractures. By explicitly representing a broad spectrum of residence times, it can distinguish rapid transport through connected channels from delayed release out of low-velocity zones. This distinction may be especially valuable when predictions must extend beyond the period covered by observations.
The implications reach well beyond a single idealized fracture. Groundwater contaminants often move through networks of intersecting fractures in which aperture, roughness, contact area, mineral coatings, and stress conditions vary from place to place. At repository sites, even a small population of fractures with large contact zones could retain dissolved radionuclides or chemical tracers and release them gradually over extended periods. In groundwater management, early arrivals through fast channels could expose wells before a monitoring network detects the full extent of a contaminant plume, while prolonged tailing could cause concentrations to remain above regulatory thresholds long after the principal plume has passed. Engineers designing geothermal systems, underground storage projects, or grouting and seepage-control measures may likewise need to account for how contact geometry modifies both permeability and transport times.
The authors emphasize that fracture geometry should be treated as a dominant control rather than a secondary correction to average flow conditions. Their results suggest that reliable prediction requires models capable of resolving, or statistically representing, the heterogeneity created by rough surfaces and direct contact. The work is based on numerical simulations, and the researchers note that data and materials will be made available on request, leaving room for future laboratory and field validation. Experiments using transparent fracture replicas, imaging of natural rock surfaces, tracer tests, and geophysical monitoring could help determine how closely the simulated contact-area effects match real systems under changing pressure and flow conditions. For now, the study provides a clear warning: underground solute transport is not governed by speed alone. In fractured rock, the places where water can barely move may be just as important as the channels where it races ahead.
Subject of Research: Solute transport in three-dimensional rough rock fractures
Article Title: Effect of flow velocity and contact area on solute transport in rock fractures
Article References: Yan, Y., Ma, L., Qian, J. et al. “Effect of flow velocity and contact area on solute transport in rock fractures.” Hydrogeology Journal (2026). https://doi.org/10.1007/s10040-026-03098-z
Image Credits: AI Generated
DOI: 10.1007/s10040-026-03098-z
Keywords: Rock fractures, contact area, flow velocity, solute transport, non-Fickian transport, continuous-time random walk, advection–dispersion equation, groundwater management, nuclear-waste disposal, fractured media
