Observable Projection and Experimental Prediction Lock for the Neutrino Carrier

Neutrino carrier ARC and HNL prediction 16/18 This paper establishes the observable projection and experimental prediction lock for the source-qualified neutrino carrier. The preceding PCR papers determine the neutrino carrier, its standing spectral structure, the joint physical target extensions, and the branch-stable one-carrier result. PCR‑7 addresses the next question: when does that structural carrier result become a genuine, experimentally testable prediction? The paper constructs a finite source-to-record protocol with fixed applicability, physical response, reconstruction, calibration, nuisance structure, uncertainty rules, and held-out comparison. The protocol is stratified because ordinary diagonal observables and predicted null directions have different mathematical behavior: diagonal Gram coordinates are reconstructed with two-sided uncertainty, while off-diagonal heavy-response coordinates are represented by squared moduli and tested using one-sided upper bounds. The preferred implementation is tagged-\(W^*\), flavor-complete charged-current spectral tomography. Its finite protocol includes: tagged charged-current parent production and source normalization; charged-lepton flavor, timing, recoil, and missing-mass calibration; light-sector spectral measurements; inclusive heavy-state production records; ordered-flavor heavy decay or scattering analyzers; pole-support reconstruction over the complete heavy interval; common carrier/reference background and control records; reconstruction of active-deficit and heavy-support matrices; rank, flavor-universality, and closure tests; and a frozen simultaneous confidence comparison with the predicted and no-vectorlike reference classes. The central observable relation is \[ \Delta_{\mathrm{act}}=\Omega_H, \] together with positive rank-three heavy support, the associated flavor structure, and heavy-pole support in the declared interval. A heavy excess without the correlated active deficit is not the predicted packet, and an active deficit without inclusive heavy support indicates a different source, loss channel, or protocol failure. A complete falsifier set is defined before held-out signal data are examined. The prediction may be falsified by failure of the active-deficit/heavy-weight equality, loss of rank-three support, absence of heavy pole support despite certified coverage, failure of flavor universality, exclusion of the carrier class in favor of the no-vectorlike reference, or failure of the stratified response and identifiability certificate. The paper specifies an eight-stage experiment map and a preregistration object containing the target child, source and analyzer cells, mass-window cover, calibration and sideband plan, exposure allocation, response model, covariance transport, confidence level, reference family, nuisance domain, and falsifier set. Later implementation may improve sensitivity but may not retune these fields without issuing a new prediction version. The result is an experimentally well-typed, finite, and falsifiable prediction framework. It does not claim that an existing experiment already realizes the complete protocol, establish present-facility sensitivity, select a unique microscopic realization, determine a numerical physical pole, or constitute empirical confirmation. A dataset that lacks the certified sensitivity threshold is indeterminate rather than automatically confirmatory or falsifying.