Cross-Domain Noise Condensation: Universal Topological Signature of Ordered-to-Disordered Phase Transitions in Typhoon, ENSO, Earthquake, and Greenland Ice Sheet

Abstract

We present the first cross-domain empirical demonstration of a universal topological signal — noise condensation (dH_curl divergence + D_fold negative) — at ordered-to-disordered phase transitions across four geophysical domains: tropical cyclones (21 WPAC typhoons), ENSO (1870-2025 SST), earthquake catalogues (USGS M≥4, 1970-2025), and Greenland ice sheet radar stratigraphy. All four systems exhibit a statistically significant collapse of noise field dimensionality (dH_curl) and a negative multifractal spectrum folding index (D_fold) at their respective phase transitions, with methodologically independent replication across multiple data products (OISST, HadISST, ERA5 for ENSO; USGS + JMA for earthquakes). The effect cannot be produced by null models (shuffle, bootstrap, spatial randomisation), establishing the noise condensation signature as a genuine structural feature of these systems. Code and data included.