Fault-zone geometry

The Geometry of Scale Dependence

Reconciling earthquake fracture energy with fractal fault-zone structures

Earthquake fracture energy is commonly normalized by an idealized planar rupture area, yet natural ruptures occupy finite-width fault zones containing gouge, subsidiary fractures, and off-fault damage. This project isolates one geometric question: how much structural information is lost when a three-dimensional fault zone is observed and measured on a two-dimensional plane?

Three-dimensional Voronoi tessellation and a two-dimensional planar slice — A numerical stereology study testing whether the reduction of a volumetric fault zone to a planar observation can contribute to the apparent scale dependence of earthquake fracture energy.

Focus: Fracture-energy scaling
Method: 3D Voronoi modelling and stereology

The scale-dependent observation

In Classic Linear Elastic Fracture Mechanics, fracture energy is defined as the energy needed to create per unit area of a new ideal plane in the material. It’s an intensive or intrinsic property of the material to quantify how hard it is to break down the atom bond to create a new surface, which means it would not change with system’s size. Earthquake-based estimates, however, span many orders of magnitude and tend to increase with slip.

Many explanations invoke scale-dependent weakening processes during rapid coseismic slip, including thermal and frictional mechanisms like flashheating. This study does not replace those mechanisms. Instead, it asks whether the geometry used to define rupture area introduces an additional, measurable bias.

Compilation of fracture energy estimates plotted against slip or critical slip distance — Compiled fracture-energy and breakdown-work estimates increase with slip across laboratory, numerical, geological, and seismological datasets. The supplied compilation draws on Cocco et al. (2023), Watanabe and Kaneko (2026), Gabriel et al. (2024), and Viesca datasets.

Rupture as a zone rather than an ideal plane

A mature fault contains a principal slip zone, a fault core, distributed gouge and cataclasite, subsidiary faults, and a damage zone extending into the wall rock. Energy can therefore be dissipated across a network of surfaces and deforming material rather than on a single mathematical plane.

If the true dissipative structure is volumetric but the inferred energy is divided by a planar rupture area, the effective area involved in failure may be underestimated which caused the scale dependence. The central objective is to quantify this dimensional-reduction effect before coupling it to more complete rupture physics.

Conceptual architecture of a natural fault core, principal slip zone, and damage zone — Conceptual fault-zone architecture showing the principal slip zone, fault core, fault gouge, and distributed damage zone. Adapted from the fault-zone framework summarized by Cocco et al. (2023).

A controlled stereological experiment

I construct a bounded, space-filling fault-gouge analogue using a three-dimensional Voronoi tessellation whose cell sizes follow a prescribed power-law distribution. The same realization is then intersected by planar sections, creating virtual outcrops that can be measured independently of the parent volume.

Because the three-dimensional body and its two-dimensional sections share the same underlying structure, differences between their recovered exponents can be attributed to dimensional sampling rather than to different geological histories or datasets.

Synthetic three-dimensional Voronoi fault-zone model and representative planar section — Synthetic 3D Voronoi fault-zone structure and a representative 2D section cut from the same realization. The experiment compares volumetric cell statistics with the statistics recovered from planar intersections.

From a 3D hierarchy to a 2D outcrop

Outcrop studies of fault gouge commonly report two-dimensional particle-size exponents of approximately 2–3, whereas the modeled three-dimensional distribution used here can require an exponent greater than 4. This contrast motivates a direct test of whether planar sampling systematically reduces the apparent exponent.

The numerical sections retain a recognizable power law but recover a lower exponent than the parent volume. For a representative realization, the volumetric estimate is D₃ᴅ = 4.511 ± 0.022 and the planar estimate is D₂ᴅ = 2.913 ± 0.012. The result shows that a convincing 2D power law can coexist with substantial loss of information about the 3D hierarchy.

Two-dimensional measurements across a mapped fault core and damage zone — Example of field-based 2D measurements across a fault core and damage zones, after Billi and Storti (2004). Such sections capture a plane through a structure that developed in three dimensions.

Comparison of a three-dimensional Voronoi tessellation with its two-dimensional slice — Dimensional sampling in the numerical model: a planar section preserves the broad scale-free pattern while compressing the particle-size hierarchy of the parent volume.

Interpretation and limits

The stereological bias provides a geometric contribution to apparent fracture-energy scaling: a planar description can undercount the distributed surface area and damage represented within a fault zone. It should be interpreted as a baseline correction, not as a complete mechanism for earthquake energetics.

The model is topological. It does not yet resolve dynamic rupture, thermal pressurization, frictional evolution, or the finite width imposed by the seismogenic zone. A next step is to couple the measured geometric bias to dynamic rupture calculations and test when the relation remains valid across scale.

Selected references

Viesca, R. C., and Garagash, D. I. (2015). Ubiquitous weakening of faults due to thermal pressurization. Nature Geoscience, 8, 875–879. https://doi.org/10.1038/ngeo2554
Cocco, M., Aretusini, S., Cornelio, C., et al. (2023). Fracture energy and breakdown work during earthquakes. Annual Review of Earth and Planetary Sciences, 51, 217–252. https://doi.org/10.1146/annurev-earth-071822-100304
Gabriel, A. A., Garagash, D. I., Palgunadi, K. H., et al. (2024). Fault size-dependent fracture energy explains multiscale seismicity and cascading earthquakes. Science, 385, eadj9587. https://doi.org/10.1126/science.adj9587
Billi, A., and Storti, F. (2004). Fractal distribution of particle size in carbonate cataclastic rocks from the core of a regional strike-slip fault zone. Tectonophysics, 384, 115–128. https://doi.org/10.1016/j.tecto.2004.03.015