Author note (2026-11-02). The paper cited here, arXiv:2603.06892 (“On the coupled geometrical-mechanical origin of the earthquake b-value in fault networks”), is by Wenbo Pan, Zixin Zhang, Bjorn Lund, Qinghua Lei. The name(s) this essay attributes it to (Pinel, Danré, Steer) do not appear among the listed authors, so the attribution below is wrong or refers to someone else's earlier work; I have not worked out which. Found by a mechanical check of each essay's named authors against the arXiv author list.
The Gutenberg-Richter law has two branches, not one.
Earthquake frequency decreases with magnitude according to a power law — the Gutenberg-Richter relation. The slope of this log-linear relationship, the b-value, is one of the most measured parameters in seismology. Typically b ≈ 1, meaning tenfold more magnitude-3 events than magnitude-4 events. But b varies across regions, depths, and tectonic settings, and the physical origin of these variations has been debated for decades.
Pinel, Danré, and Steer (arXiv:2603.06892) show that the b-value emerges from the coupling of two scaling laws: fault rupture area scales as a power law with fault size, and slip magnitude scales separately with the mechanical conditions. These two scalings produce not one but two distinct branches in the frequency-magnitude relationship.
The first branch describes earthquakes where rupture fills the entire fault — the event size is set by geometry (how big is the fault?). The second branch describes earthquakes where rupture stops before reaching the fault's edges — the event size is set by mechanics (how much stress was available?). The transition between branches occurs at a characteristic magnitude determined by fault criticality and the number of faults activated during mainshock-aftershock sequences.
The standard analysis fits a single line through both branches and reports one b-value. This averaging obscures the transition and produces apparent b-value variations that are actually shifts in which branch dominates the data. A region with many small faults (geometry-limited) has a different apparent b-value than a region with few large faults (mechanics-limited), even if the underlying physics is identical.
Geometry and mechanics, coupled but separable. The statistic that appears to be one parameter is actually two parameters measured through one number.
Pinel, Danré, and Steer, "On the coupled geometrical-mechanical origin of the earthquake b-value in fault networks," arXiv:2603.06892 (2026).