Fragment-Asperity Model in Seismology

• The Fragment-Asperity Model is a seismomechanical framework representing faults as ensembles of stress-bearing asperities and gouge fragments that trigger seismic events.
• It employs nonextensive Tsallis entropy to derive fragment area distributions and establishes a unique surface-energy scaling linking released energy to contact surface area.
• Model distinctions between fragment breakage and asperity rupture provide insights into preseismic electromagnetic emissions and Gutenberg-Richter scaling of seismicity.

Searching arXiv for papers on the Fragment-Asperity Model and closely related asperity-based frameworks. The Fragment-Asperity Model is a seismomechanical framework in which an earthquake fault is represented as an ensemble of asperities—stress-bearing surface contacts—and fragments—chips or gouge material between sliding blocks. In this picture, seismic events occur when asperities break or fragments shift, releasing strain. Within the nonextensive statistical formulation of the model, earthquake populations are described with Tsallis entropy, and a recent development identifies a linear relation between released energy and fragment or asperity contact surface as the unique scaling that yields a closed elementary expression for the total entropy as a function of the non-extensivity parameter $q$ (Sotolongo-Costa et al., 16 Mar 2026). Earlier work had already placed the model in a broader nonextensive description of earthquake dynamics, alongside a distinct self-affine asperity model, and used it to interpret preseismic electromagnetic emissions (Minadakis et al., 2011).

1. Physical picture and model lineage

The model treats a fault zone as two rough, irregular surfaces with a gouge of fragments between them. Its basic entities are therefore not only the fault-plane irregularities themselves, but also the intervening fragmented material. The triggering mechanism is established through the interaction of the irregularities of the fault planes and the fragments between them, so that rupture may be associated either with the fracture of fragments in the gap or with the fracture of larger load-bearing asperities, described as “teeth” on the fault surfaces (Minadakis et al., 2011).

This distinction is important because the Fragment-Asperity Model is not identical to the self-affine asperity model. In the self-affine model, an earthquake is due to the slipping of two rough and rigid fractional Brownian profiles when there is an intersection between them; in the fragment-asperity approach, the gap-filling fragments are constitutive elements of the dynamics. A plausible implication is that the fragment-asperity formulation is better suited to fault gouge environments in which granular and fracture processes are inseparable from contact mechanics.

2. Nonextensive statistical formulation

The statistical mechanics of the model are expressed with Tsallis entropy, which is used for systems with long-range interactions, memory effects, and fractal or multifractal organization. In discrete form, the entropy is written as

$$S_q = k \frac{1}{q - 1} \left(1 - \sum_{i=1}^{W} p_i^q \right),$$

with the Boltzmann-Gibbs limit recovered at $q=1$ (Minadakis et al., 2011). In the 2026 development, the continuous entropy for the fragment surface distribution is written, for $k=1$, as

$$S = \frac{1 - \int_{0}^{\infty} p^q(\sigma)\, d\sigma}{q-1},$$

where $p(\sigma)$ is the probability distribution of fragment surface area $\sigma$ (Sotolongo-Costa et al., 16 Mar 2026).

Maximization of the entropy with constraints yields the fragment area distribution

$$p(\sigma) = (2-q)^{1/(2-q)} \left[1 + (q-1)(2-q)^{(q-1)/(2-q)} \sigma \right]^{1/(1-q)}.$$

In this formulation, $q>1$ characterizes non-extensive, long-range correlated, multifractal systems, which matches the phenomenology assigned to seismicity. Earlier nonextensive fits in seismic and electromagnetic catalogues reported $q\sim 1.6-1.8$, while the recent entropy-based analysis identifies a broader critical interval $$S_q = k \frac{1}{q - 1} \left(1 - \sum_{i=1}^{W} p_i^q \right),$$0 for main shocks [(Minadakis et al., 2011); (Sotolongo-Costa et al., 16 Mar 2026)].

3. Rupture modes and electromagnetic precursors

A major interpretive use of the model has been the discrimination between two rupture modes. Fragment fracture corresponds to the breaking of small to medium fragments in the fault gouge; asperity or “teeth” fracture corresponds to the rupture of large, strong protrusions on the fault surfaces. The former is associated with lower energies and weaker organization, whereas the latter is associated with sudden, larger energy release and the final stage of earthquake generation (Minadakis et al., 2011).

This distinction was linked to preseismic kHz electromagnetic emissions. In the cited two-stage interpretation, sparse, weakly organized, anti-persistent emissions with $$S_q = k \frac{1}{q - 1} \left(1 - \sum_{i=1}^{W} p_i^q \right),$$1 are associated with fragment breakage, while abrupt, densely clustered, persistent emissions with $$S_q = k \frac{1}{q - 1} \left(1 - \sum_{i=1}^{W} p_i^q \right),$$2 are associated with asperity rupture. The same study also reports lower Tsallis entropy during the second epoch, indicating increased organization. This suggests that the Fragment-Asperity Model is not only a magnitude-distribution framework but also a dynamical scheme for classifying stages of rupture preparation (Minadakis et al., 2011).

4. Surface-energy scaling and entropy uniqueness

The principal theoretical advance of the 2026 paper is the proposal and proof that the released energy $$S_q = k \frac{1}{q - 1} \left(1 - \sum_{i=1}^{W} p_i^q \right),$$3 is proportional to the surface area $$S_q = k \frac{1}{q - 1} \left(1 - \sum_{i=1}^{W} p_i^q \right),$$4 of contact fragments and asperities: $$S_q = k \frac{1}{q - 1} \left(1 - \sum_{i=1}^{W} p_i^q \right),$$5 This replaces previous assumptions in which $$S_q = k \frac{1}{q - 1} \left(1 - \sum_{i=1}^{W} p_i^q \right),$$6 or $$S_q = k \frac{1}{q - 1} \left(1 - \sum_{i=1}^{W} p_i^q \right),$$7, where $$S_q = k \frac{1}{q - 1} \left(1 - \sum_{i=1}^{W} p_i^q \right),$$8 denotes fragment size. The physical argument is drawn from Griffith fracture theory: the energy required to fracture material is linked directly to the creation of new surface area rather than to volume (Sotolongo-Costa et al., 16 Mar 2026).

The same work argues that this surface-energy scaling is mathematically unique within the model because it is the only one that yields a closed elementary form for a well-defined total entropy as a function of $$S_q = k \frac{1}{q - 1} \left(1 - \sum_{i=1}^{W} p_i^q \right),$$9, without ad hoc cutoffs or special functions. If the energy scales differently with $q=1$0, the entropy integral either diverges or cannot be written in closed form. This gives the scaling assumption a status stronger than a convenient ansatz: it becomes the condition under which the nonextensive thermodynamic description remains analytically closed (Sotolongo-Costa et al., 16 Mar 2026).

5. Criticality, the $q=1$1-interval, and the Gutenberg-Richter connection

By writing the total Tsallis entropy as a function of $q=1$2, the 2026 analysis identifies a critical range

$q=1$3

with maximum entropy near $q=1$4. The same interval contains most reported $q=1$5-values determined for main shocks around the world in recent decades, including tectonic regions such as Mexico, Ecuador, Italy, and the San Andreas system. The entropy varies steeply within this interval, and empirically observed $q=1$6-values fall where the rate of change of entropy with $q=1$7 is greatest (Sotolongo-Costa et al., 16 Mar 2026).

The paper therefore interprets $q=1$8 not merely as a fitting parameter but as a criticality indicator. It further notes detectable precursory behavior: $q=1$9 often increases before large main shocks and then drops after, with examples including the 2016 Ecuador Manabí earthquake, the 1995 Kobe event, and the 2009 L’Aquila event. This places $k=1$0 in a role analogous to a phase or criticality marker in thermodynamic transitions (Sotolongo-Costa et al., 16 Mar 2026).

The model is also connected to the Gutenberg-Richter law,

$k=1$1

through the energy-magnitude relation $k=1$2. Comparison of the non-extensive cumulative energy distribution,

$k=1$3

with the Gutenberg-Richter form,

$k=1$4

yields

$k=1$5

The paper emphasizes the $k=1$6 factor as a consequence of the surface-driven energy scaling together with volume-based event counting (Sotolongo-Costa et al., 16 Mar 2026).

6. Related asperity-based frameworks and modeling boundaries

The Fragment-Asperity Model belongs to a broader family of asperity-based descriptions, but these adjacent frameworks address different physical problems and impose different assumptions. In rough-contact tribology, one recent asperity-based statistical model treats individual junctions with JKR theory under normal loading and the Papangelo–Ciavarella model under mixed-mode loading, while integrating over a Gaussian asperity-height distribution. That model reproduces a static friction peak and shear-induced reduction in real contact area, and explicitly places itself within the broader Fragment-Asperity Model framework, including Greenwood-Williamson and Fuller-Tabor lines of development (Xu et al., 2022).

In contact mechanics more generally, asperity models remain sensitive to statistical assumptions. A comparative study of Gaussian and Weibull summit-height distributions reports that area-load linearity is robust, whereas load-separation and stiffness-load relations can change strongly when Gaussianity is relaxed. The practical conclusion is that Gaussian height statistics should be experimentally tested before using sophisticated rough-contact models built on that assumption (Ciavarella, 2016).

In subsurface fracture mechanics, a non-local closure model parameterizes roughness and asperity deformation with a contact law involving $k=1$7 and $k=1$8, and captures progressive, non-uniform closure from fracture edges toward the center (Wang et al., 2017). In brittle ceramics, fragment statistics extracted from crack coalescence are used to define a transition from cracked solid to granular medium through the Effective Fragmentation Ratio, with onset proposed at $k=1$9–$$S = \frac{1 - \int_{0}^{\infty} p^q(\sigma)\, d\sigma}{q-1},$$0 (Bhattacharjee et al., 2020). In triboelectricity, an asperity model incorporating sphere, cylinder, cone, roller, and punch geometries, together with gradient elasticity, shows that electromechanical response depends strongly on asperity shape and nanoscale size effects; that paper explicitly states that area-based or spherical-asperity representations miss important scaling behavior (Olson et al., 2022).

Domain	Representative paper	Relevance to fragment-asperity thinking
Seismicity and EM precursors	(Minadakis et al., 2011)	Distinguishes fragment fracture from asperity fracture
Adhesive rough friction	(Xu et al., 2022)	Statistical asperity population with adhesion-friction coupling
Rough-contact statistics	(Ciavarella, 2016)	Sensitivity to height-distribution assumptions
Fracture closure	(Wang et al., 2017)	Non-local asperity contact and progressive closure
Granular transition	(Bhattacharjee et al., 2020)	Fragment statistics and transition criteria
Triboelectric contacts	(Olson et al., 2022)	Shape and gradient-elasticity effects at asperities

These neighboring literatures clarify a common misconception: “asperity model” is not a single formalism. The seismic Fragment-Asperity Model is specifically a nonextensive earthquake model in which fragments in the gouge and asperities on the fault surfaces jointly control rupture statistics and criticality. Related asperity-based models in tribology, fracture closure, granular transition, and triboelectricity share the language of rough contacts and local junction mechanics, but they solve different constitutive and observational problems.