QADI Prediction and Falsification Registry: Protected Equations, Provenance, ζ Audits, and Replication Protocols

The QADI Prediction and Falsification Registry, Version 1.0, establishes a frozen, auditable set of 23 quantitative or discrete predictions derived from the Aether Physics Model (APM) and Quantum Measurement Units (QMU). The registry is maintained by the Quantum AetherDynamics Institute (QADI) as a preregistration and falsification ledger. Registration is explicitly distinguished from empirical validation: inclusion of a prediction records the claim to be tested and does not constitute evidence that the prediction is correct. Each protected prediction is assigned a permanent identifier, QADI-P001 through QADI-P023, and records its source publication, protected mathematical or discrete statement, provenance, observable, comparator or null hypothesis, decision rule, current state, ζ-invariance status, and execution class. Source publications are linked directly through their Zenodo DOIs. The registry is designed to prevent post hoc alteration of predictions after comparison with experimental or observational data. Once a protected prediction has been confronted with its designated data, its historical mathematical statement is retained. A substantive change to the protected equation, fitted parameters, data selection, null model, or pass/fail criterion requires a new registry entry or an explicitly superseding registry version. The empirical-primitives audit applies the transformation mₐ↦ζ mₐ,ₐ²↦ζ eₐ²,αₐ↦αₐ/ζ,↦ζ\,Gforce,/ζ, and identifies which registered observables remain invariant under the empirical primitive convention. The registry also distinguishes three execution classes: Class A --- predictions executable using external observational or archival data; Class B --- predictions executable using conventional laboratory methods; Class C --- predictions requiring dedicated apparatus derived from the APM/QMU experimental program. Retrospective evidence is explicitly separated from prospective replication or holdout tests. Agreement with previously analyzed data does not substitute for execution of a protected prospective test. Several independent predictions share the topological factor 1/16π², including the registered cosmic-birefringence angle, Casimir coefficient, and Aether-unit stiffness ratio. The registry does not count recurrence of this factor as independent empirical corroboration unless the quantities are obtained through genuinely independent physical realizations without a shared fitted scale. The Casimir entry explicitly distinguishes the conventional electromagnetic coefficient written in an h-based convention, π²/240=π h/480, from the APM prediction CAPM=1/16π². The registry includes predictions involving cosmic polarization rotation, Casimir boundary stress, nonlinear phonon helicity transfer, Tesla-resonator charge partition and geometric phase, Aether-unit metrology, directional curl-gradient interferometry, astrophysical torsion filaments, thermoresonant tellurium, high-Z hydrogenic spectroscopy, charged-lepton magnetic anomalies, localized gravity tests, Aether displacement-wave polarization, photoemission, plasma transport, optical chirality, neutron lifetime measurements, and one-spin phtn closure. Version 1.0 is intended to serve as the fixed baseline against which subsequent experimental, observational, and replication results can be scored. Null, failed, and inconclusive results remain scientifically relevant and are to be reported against the originally registered statement rather than used to redefine it. The central methodological distinction is registration status validation status. The purpose of this registry is therefore not to assert that 23 predictions have been experimentally confirmed, but to establish in advance exactly what APM/QMU predicts and what observations would count against those predictions.