Pre-Seismic Magnitude Compression Across Six Tectonic Regimes: Evidence from K–R Critical Slowing-Down Indicators in Nine Earthquake Catalogs

We introduce the K–R excitation–regulation framework, a coupled ordinary differential equation (ODE) system that produces critical slowing-down (CSD) indicators from earthquake magnitude sequences. The K–R ODE integrates excitation K and regulation R event-by-event: dK/dt = α tanh(σ) − βK, dR/dt = γK − δR, where σ = (M − Mc)/(9.5 − Mc). An identical eight-step pipeline is applied to nine independent catalogs spanning six tectonic regimes: Japan and Chile (subduction), San Andreas (strike-slip), New Zealand (oblique subduction), Greece (back-arc extension), Sumatra (Sunda subduction), Turkey (continental collision), ISC-GEM Global (35,748 events), and Central-Eastern US (intraplate null control). CSD₅₀, the 50-event rolling standard deviation of K, is tested at five pre-seismic lags (−21 to −1 days) under three temporal isolation criteria (30, 60, 90 days) with Benjamini–Hochberg false-discovery-rate correction. Eight of nine catalogs show systematic suppression of 6–18% before large mainshocks. Across all catalogs, 68 of 127 testable combinations achieve FDR significance, substantially exceeding the null expectation. San Andreas achieves FDR significance in all tested lag-isolation combinations. Sixteen ETAS simulations establish that rate reduction alone produces no suppression (−1.5%, p = 0.89), whereas magnitude compression reproduces the observed effect (−16.1%). Magnitude compression is statistically independent of rate quiescence (r = −0.009). Central-Eastern US correctly returns no signal. These results identify a previously underexplored pre-seismic pattern—magnitude distribution compression—that is consistent across plate-boundary regimes and should be validated prospectively.