Role of plate convergence rate in shaping earthquake recurrence in subduction zones.

Understanding the complex interplay of subduction zone processes is key to unravelling the timing and distribution of great earthquake cycles within the framework of the plate tectonics paradigm. Megathrust earthquakes, though extensively investigated and their quasi-repetitive nature well recognised, remain challenging to assess globally due to their long recurrence intervals and limited historical data. Slow earthquakes in the brittle-to-ductile transition zone interact dynamically with megathrust events, but their potential to trigger or delay large earthquakes remains unclear. The periodic nature of slow earthquakes (typically recurring over timescales of months to years) has enabled detailed seismic and geodetic catalogs, offering new insights into subduction zone dynamics. Here, we adopt a tripartite approach, integrating natural observations, numerical simulations, and laboratory experiments to investigate relationship between megathrust and slow earthquakes in subduction zones. Analysis of tremor catalogs of Cascadia and Nankai subduction zones, reveal a systematic logarithmic inverse relationship between recurrence intervals (T) and plate convergence rates (V), with downdip tremor patches exhibiting shorter recurrence times than updip segments. Quasidynamic rate-and-state friction (RSF) simulations, calibrated against geodetic displacements, demonstrate that this periodicity arises from frictional healing modulated by V (T ∞ -V). Laboratory stick-slip experiments validate this scaling, showing force drop and recurrence times decrease logarithmically with increasing loading velocity, consistent with natural and numerical observations. Crucially, the logarithmic dependency persists across tectonic (months-years), numerical (stick-slip cycles), and laboratory (seconds-minutes) scales, resolving ambiguities in scaling fault mechanics from lab to tectonic regimes. We propose that SSEs in the brittle-ductile transition zone episodically transfer stress to adjacent locked megathrust segments, acting as real-time 'stress-meters' for seismogenic zone dynamics. This framework bridges geophysical monitoring, Rate and State Friction theory, and experimental fault mechanics, offering a predictive tool to infer stress accumulation on megathrusts. By linking slow earthquake periodicity to plate kinematics, our findings advance a unified paradigm for earthquake cycle dynamics, directly informing probabilistic hazard models and mitigation strategies in subduction zones. The integration of multi-scale constraints underscores the potential of tremor networks to monitor stress evolution, enhancing our capacity to identify regions at risk of large seismic events.