Astronomical cycles have long been proposed as potential modulators of seismicity, yet their influence remains controversial due to inconsistent results and methodological limitations. In this study it was tested whether the 18.6-year lunar nodal cycle, expressed in Earth rotation through length-of-day (LOD) variations, is detectable in global seismicity using independent earthquake catalogs (ISC–GEM, GCMT, USGS/NEIC) and a fully reproducible statistical framework. Monthly time series of earthquake counts and cumulative seismic moment release were analyzed using cross-spectral methods and generalized linear models that explicitly account for non-stationarity. Cross-spectral analysis reveals frequency-specific behavior, with no significant coherence between LOD and earthquake counts at the 18.6-year period, but intermittent and statistically significant coherence with seismic moment release (p 0.05 in up to ∼40% of windows). Likelihood-based models show that inclusion of the 18.6-year component improves the description of seismic moment release (ΔAIC ≈ −10 to −15; deviance reduction ≈ 5–8%), corresponding to a modulation amplitude of approximately 8%–12% between nodal extrema. These improvements are robust across catalogs and retained in out-of-sample validation. A latitudinal dependence of seismicity is observed, consistent with previously proposed astronomical tuning on plate tectonics. This is also supported by the evident decrease of the seismic moment moving toward both either the geographic or tectonic poles. Spatial analysis further reveals a hemispheric asymmetry in seismic energy release when expressed in a tectonic reference frame, consistent across reasonable variations in pole definition. These results indicate that the 18.6-year nodal cycle does not modulate earthquake occurrence rates but may modulate the energetic expression of global seismicity. The findings support a view of the lithosphere as a nonlinear system in which astronomical forcing can influence energy release without deterministically triggering earthquakes.
Carlo Doglioni (Wed,) studied this question.