Rethinking Earthquakes

Abstract. Earthquake science has progressed through major paradigm shifts regarding how we measure and conceptualize the size and physics of earthquakes. While seismology prior to the 1950s viewed earthquakes as dynamic, wave-dominated processes quantified by radiated energy, the establishment of plate tectonics in the 1960s shifted the focus entirely to macroscopic fault motions measured by the seismic moment. This fault-centric view, combined with laboratory-derived rate- and state-dependent friction (RSF) laws, fostered an optimism during the late 20th century that the entire earthquake process—including deterministic nucleation and precursors—could be comprehensively explained by continuum mechanics. However, the 21st-century discovery of slow earthquakes has fundamentally challenged this unified picture. This review rethinks the governing physics of both fast and slow earthquakes by contrasting their dynamic behaviours. Slow earthquakes, characterized by a linear moment–duration scaling, scale-dependent migration velocities, and Brownian fluctuations, are fundamentally governed by diffusion processes rather than the wave equation. Conversely, fast earthquakes represent a scale-free, stochastic cascade-up process on hierarchical fault structures. Decades of high-resolution observations indicate that macroscopic, deterministic nucleation phases are practically negligible, and that the earthquake growth process is well-approximated by a single, universal moment growth curve. Within this framework, the Gutenberg–Richter law can be reinterpreted not merely as a regional statistic, but as the conditional probability of a cascading rupture continuing or terminating, which is modulated by local environmental conditions such as stress and fault geometry. Ultimately, shifting our perspective to this stochastic, dual-mode nature of fault motion—powered by modern dense networks, Artificial Intelligence, and high-performance computing—provides a new foundation to unlock the remaining mysteries of fault mechanics and earthquake predictability in the coming decades.