When Is a Time Field Observable? Identifiability and Null Tests with Clocks, Gravity, and Radio Transients

This methodological synthesis asks when a scalar interpretation of timing residuals is distinguishable from reference choices and conventional effects. A covariant low-energy action, an explicit Jordan--Einstein normalization map, and matched clock responses connect field parameters to observables. For fixed linear Gaussian means, identifiable combinations are precisely the non-null directions of the whitened, nuisance-projected response. A published Sr--Yb+--Cs sensitivity matrix has rank two for three coupling--amplitude products; additional epochs do not remove its null direction. In 20,000 synthetic realizations, nominal 95% marginal intervals attain 95.10% and 95.09% coverage with the full covariance, but only 71.99% and 72.37% when temporal correlation is omitted. Separate covariance ablations identify the origin of this failure. Additional full-forward checks include gaps, nonseparable noise, unknown phase and an independently calibrated frequency scan with estimated noise scale. A retrospective diagnostic of 25 public GNSS clock products verifies 7,072 common increments and exact contrast closure; it is not a scalar-coupling measurement. Cassini and MICROSCOPE summaries supply conditional restrictions with explicit normalization and applicability. The minimal electromagnetic coupling preserves the geometric-optics vacuum light cone; both achromatic and ν⁻² extra single-burst delays are absorbed by unrestricted conventional templates. These worked calculations support falsifiable search design, not a detected time field, a new statistical law or a demonstrated unification. Files and languages: the English PDF is the main manuscript; the Spanish PDF is a translation of the same study, not a separate result. fuentes_bilingues.zip contains the editable LaTeX sources and article license. reproducibilidad_publica.zip contains both sources, executable checks, fixed protocol, numerical outputs, a manifest with 25 original IGS product URLs and SHA-256 hashes, and a verified downloader. Raw third-party products and private project reports are not redistributed. Status and scope: methodological preprint, not independently peer reviewed. Synthetic tests are not observations. The GNSS calculation is a retrospective engineering diagnostic; no scalar-field detection, new experimental scalar-coupling bound, or achieved unification is claimed. Related prior publication: Roberto Ros Velasco (2025), Temporal-phase and fractal-decoherence extensions for JUNO oscillation analyses — The Equation of Time framework, https://doi.org/10.5281/zenodo.17198769. That two-page neutrino proposal is a distinct work; this record is not a new version of it. Licensing is component-specific: article, Spanish translation and original figures, CC BY 4.0; original code, MIT. These are not alternative licenses covering every component. Third-party sources retain their original rights and attribution. No external funding. No competing interests declared by the author. Generative-AI assistance is disclosed in both manuscripts; the author retains scientific responsibility. Correspondence: arquitec.robertoros@gmail.com.