A geometrical aperture-width relationship for rock fractures

The relationship between fracture aperture (maximum opening; dmax) and fracture width (w) has been the subject of debate over the past several decades. An empirical power law has been commonly applied to relate these two parameters. Its exponent (n) is generally determined by fitting the power-law function to experimental observations measured at various scales. Invoking concepts from fractal geometry we theoretically show, as a first- order approximation, that the fracture aperture should be a linear function of its width, meaning that n = 1. This finding is in agreement with the result of linear elastic fracture mechanics (LEFM) theory. We compare the model predictions with experimental observations available in the literature. This comparison generally supports a linear relationship between fracture aperture and fracture width, although there exists considerable scatter in the data. We also discuss the limitations of the proposed model, and its potential application to the prediction of flow and transport in fractures. Based on more than 170 experimental observations from the literature, we show that such a linear relationship, in combination with the cubic law, is able to scale flow rate with fracture aperture over ~14 orders of magnitude for variations in flow rate and ~5 orders of magnitude for variations in fracture width.

💡 Research Summary

The paper addresses the long‑standing problem of how the maximum aperture (dₘₐₓ) of a rock fracture scales with its width (w). While many field and laboratory studies have fitted an empirical power law dₘₐₓ = c wⁿ, the exponent n has varied widely (from about 0.5 to over 2) depending on the dataset, measurement technique, and scale. Linear elastic fracture mechanics (LEFM) predicts n = 1, but empirical evidence has been ambiguous.

The authors propose a first‑order theoretical model based on fractal geometry. They idealize a single fracture as an ellipse whose boundary is rough and fractal. Using Mandelbrot’s relationship between the perimeter P and area A of a fractal object (P ∝ A^{D_b/2}, where D_b is the boundary fractal dimension, 1 ≤ D_b < 2) and the scaling law for a fractal length L_f (L_f ∝ R^{1‑D_l} L_s, with measurement scale R, straight‑line distance L_s = w, and D_l ≈ D_b), they derive that the perimeter is directly proportional to the fracture width when the measurement scale tends to zero.

Assuming the area of the rough‑boundary ellipse can be approximated by A ≈ π (dₘₐₓ/2)(w/2), substitution into the perimeter‑area relation yields a simple linear relationship:

 dₘₐₓ = C w

where C is a numerical prefactor that depends on the (very small) measurement scale but not on the fractal dimension. In other words, the aperture‑width scaling is governed by geometry rather than surface roughness; the roughness exponent cancels out. This result coincides with the LEFM prediction that aperture scales linearly with crack length, where the proportionality constant would involve material properties (effective stress, Poisson’s ratio, shear modulus).

To test the model, the authors re‑examine two datasets from Hatton et al. (2014) collected in the Krafla fissure swarm (Kelduhverfi and Myvatn). Both datasets show a large scatter of aperture values for a given width, spanning several orders of magnitude. By plotting the theoretical line dₘₐₓ = C w with three representative C values (0.001, 0.01, 0.1), they demonstrate that the envelope of these lines captures the overall trend of the data despite the scatter. The same approach successfully envelopes data from 14 independent experimental studies comprising more than 170 fractures, confirming the robustness of the linear scaling across a wide range of rock types and scales.

Combining the linear dₘₐₓ‑w relationship with the classic cubic law for flow in a single fracture (Q ∝ w d³) yields a scaling for the flow rate Q that is proportional to w (dₘₐₓ)³. Using the empirical data, the authors show that this combined model can predict flow rates over roughly 14 orders of magnitude and fracture widths over about 5 orders of magnitude, far exceeding the range of any single laboratory experiment.

The paper also discusses limitations. First, real fractures are only approximately fractal; the fractal dimension may vary with the observation scale, and the assumption of an infinitesimally small measurement scale is an idealization. Second, the elliptical geometry is a simplification; natural fractures can be highly irregular, intersecting, or have non‑elliptical shapes, which could introduce deviations from linearity. Third, measurement errors, especially in field settings where aperture is inferred from tape measurements, contribute to the observed scatter and to the variability of fitted n values reported in the literature.

Despite these caveats, the authors argue that the linear aperture‑width relationship provides a parsimonious, physically‑based foundation for upscaling fracture hydraulic properties. It bridges fractal geometry with classical fracture mechanics, offering a unified framework that can be incorporated into discrete fracture network (DFN) models and larger‑scale reservoir simulations. Future work is suggested to (i) extend the analysis to multi‑fracture systems, (ii) explore the impact of varying fractal dimensions across scales, and (iii) integrate high‑resolution imaging (e.g., X‑ray CT) to better quantify the measurement scale dependence of C.

In summary, the study demonstrates that, when viewed through the lens of fractal geometry, the maximum aperture of a rock fracture scales linearly with its width (n = 1). This simple relationship, validated against extensive experimental data, reconciles divergent empirical findings and provides a solid basis for predicting fluid flow and transport in fractured rock masses.