SCFL Operators Test and Validation Notebook: Cross-Domain Single-Event Verification, Adversarial Null-Testing Protocol, and Gap-4 Brittle-Seam Generalization (Grid, Seismic, Volcanic, and Nuclear Cascade Domains)

This notebook consolidates single-event testing of the SCFL operator family (Rupture, Shift, Drift, Stress, Kinetic, Temporal Coherence, Migration Coherence) across nine real-world and synthetic events spanning five domains: grid (ERCOT Winter Storm Uri, Winter Storm Elliott across four balancing authorities), seismic (Ridgecrest, Northridge, Tohoku, Kumamoto), volcanic (Kilauea 2018), nuclear/industrial (Fukushima Daiichi), and synthetic grid topology (ACTIVSg2000, ACTIVSg10k). Two structural contributions define this session. First, a formal adversarial null-testing protocol was applied retroactively across the operator family, reversing two prior conclusions — strengthening the Uri Temporal Coherence (Quench) result beyond its original confidence, and retracting an apparent four-domain seismic Temporal Coherence finding as a rate-of-change artifact indistinguishable from a matched Poisson null. Second, the Gap-4 seam-typing analysis identified and mathematically generalized a “brittle seam” failure signature: bridge-isolated subgraphs show DC power-flow propagation flat at numerical noise across all graded admittance reduction, followed by discontinuous, undefined failure at total fragmentation — a pattern proven load-independent by conservation of power and confirmed as reproducing across two independent synthetic networks (ACTIVSg2000 and ACTIVSg10k, differing by network size, geography, and topology). The Fukushima Daiichi cascade was reconstructed at 5-minute resolution, correcting an earlier account that conflated the earthquake and tsunami into one event; verified TEPCO timing shows the tsunami, not the earthquake, as the operative external forcing event, with a genuine ~45-minute calm plateau between the two. All results retain single-event status: no operator in this notebook has been cross-event verified, and “Locked” denotes cross-source verification for one event only (e.g., Uri cross-checked against EIA-930), not generalization. Deterministic/near-deterministic results (the Stress discriminator, the ACTIVSg2000/ACTIVSg10k brittle-seam finding) are flagged as not yet adversarially null-tested in the same statistical sense as ACF-based results, since a matched-random-noise null does not directly apply to them. Retracted results (seismic Temporal Coherence, the KBHM tornado barometric pressure signature, an early Uri Drift claim, an early ERCOT/ACTIVSg2000 variance-spike claim) are retained for the record rather than removed.