Decoding Earthquake Cycles: Plate convergence rates shape recurrence intervals in Subduction Zones

Megathrust earthquakes at subduction plate interfaces have been extensively investigated, with their quasi-repetitive nature well recognized, yet their long return periods and sparse historical records complicate global assessments of this regularity. Slow earthquakes occurring in the brittle-to-ductile transition zone demonstrate a complex interplay with large subduction earthquakes, though their roles in triggering or delaying significant events remain poorly understood. The periodicity of slow earthquakes, characterized by recurrence intervals ranging from months to years, has facilitated the creation of comprehensive seismic and geodetic event catalogues. Here, we investigate the behaviour of slow earthquakes and megathrust ruptures using integrated constraints from natural observations, numerical simulations under the rate and state friction model and laboratory-based experimental results. Focusing on the best instrumentally monitored and mature subduction zones, namely, Cascadia and Nankai, we identified a depth-dependent pattern in slip periodicity and a corresponding increase in cumulative tremor counts downdip from the trench. Our numerical simulations suggest a logarithmic dependency between recurrence time and loading velocity, consistent with the depth dependency of the tremor activities and associated slip-periodicity observed in these subduction zones. Moreover, the long-term aseismic slip distribution patterns of these subduction zones match with the model-predicted displacements for the corresponding loading velocities, which never exceed the down-dip plate motion at these subduction zones. Laboratory experimental results validate the link between recurrence time and loading velocity, establishing a connection between recurrence time and force drop as well. Further, analysis of seismic data of slow and megathrust earthquakes across major subduction zones worldwide underscores a consistent logarithmic inverse relationship between the recurrence times of these events and plate convergence rates. Our numerical simulation results and stick-slip laboratory experiment observations complement the naturally observed logarithmic behaviour of both megathrust and slow earthquakes. Integrating these insights from natural observations, numerical modelling, and experimental data, we finally argue a possible stress transfer mechanism from the slow earthquakes source zone to the adjacent megathrust earthquake segments and suggest that the slow earthquakes can be used as a possible proxy or “stress-meters” for large megathrust earthquakes and probably modulate the megathrust earthquakes in the seismogenic zone. Understanding the interplay between slow and megathrust earthquakes is crucial for seismic hazard assessment and enhances our ability to identify regions at risk of large seismic events and improve mitigation strategies.