Reinterpreting the Recurrence Time–Moment Relation of Global Repeating Earthquakes: A Scattered Data Cloud Bounded by Crack‐Model Physics

Repeating earthquakes are recurrent ruptures of the same asperity along creeping faults, and their recurrence time–moment (–) scaling provides key constraints on earthquake physics and fault creep. However, previous studies have reported , seemingly inconsistent with the classic crack model that predicts . Compiling expanded global repeater catalogs, we find a highly scattered – distribution that precludes a meaningful global linear fit. Rather than fitting a single exponent to the scattered data, we reframe the scaling problem as explaining the structure of the entire – distribution. Focusing first on the creeping San Andreas Fault, we select long‐term quasi‐periodic repeaters and correct for differences in background creep rate. These procedures yield a triangular – distribution with a robust lower envelope of slope 1/3 that defines a physically meaningful bound consistent with asperity failure by total long‐term steady fault creep, together with numerous long‐recurrence sequences above this lower bound, which likely reflect (1) higher stress drops, and/or (2) local creep‐rate variations associated with complex fault structures, and/or (3) asperity interactions. Applying the same selection and correction procedures to global catalogs reveals a similar bounded distribution. This physical alternative interpretation for the observed bounded distribution makes a falsifiable prediction: with extended monitoring and improved catalogs, additional repeater sequences should populate longer recurrence times without systematically crossing the physical 1/3 lower bound. Our reinterpretation supports the applicability of the crack model for estimating slip and highlights the potential of repeaters as high‐resolution creep meters.