Before a large earthquake, the smaller earthquakes get quieter. Not fewer — quieter. The variance of seismic activity drops.
Pasupuleti analyzed earthquake catalogs from Japan and Chile using a critical slowing down framework and found the same signal in both: a 17-21% suppression of excitation-regulation indicators across the four pre-seismic time lags before clean M >= 6.0 mainshocks. The effect isn't a reduction in earthquake frequency — it's a reduction in the variability of earthquake patterns. The fault system becomes more regular before it fails.
Note (2026-11-09): I checked the abstract of Pasupuleti's paper. It reports that a derived indicator, CSD50, is suppressed by about 17 to 21 percent before M >= 6.0 mainshocks in the Japan catalog (41 events) and about 18 to 22 percent in the Chile catalog (58 events). Its synthetic tests say variance reduction alone produces strong suppression while rate reduction alone does not, and that a combined rate-and-variance scenario matches the observations, so the essay's statement that the earthquakes get quieter, 'not fewer', is stronger than that. The abstract says 'we do not claim spatial universality or deterministic prediction' and gives a forecasting AUC of 0.549; the essay's statement that the signal reflects how faults approach failure in general goes beyond it.
This is the signature of critical slowing down. Complex systems approaching a transition lose their ability to recover from perturbations, which paradoxically makes their short-term behavior more predictable. The fluctuations that characterize a healthy, subcritical system quiet down as the system approaches its critical point. The quiet is the warning.
The cross-catalog replication is the strongest evidence. Japan and Chile are different subduction zones with different geometries, different plate velocities, different seismic histories. The same suppression pattern appearing independently in both — with statistically indistinguishable effect sizes — suggests the signal is structural rather than regional. It reflects something about how faults approach failure in general, not how Japanese or Chilean faults specifically behave.
The author is careful to note that this is not deterministic prediction. The signal doesn't tell you when or where the earthquake will occur. It tells you that the system's dynamics have changed. The quiet is detectable but not actionable — not yet.
Critical slowing down was originally formalized in ecology and climate science. Finding it in earthquake catalogs extends the principle: systems approaching tipping points lose variability before they lose stability.