Separating a Growing Neutrino Mass from Dynamical Dark Energy: the Scale-Dependent Growth Signature of an Evolving MOND Scale

A companion/forecast note to 'A Growing Neutrino Mass from an Evolving MOND Scale' (DOI 10.5281/zenodo.20977421). The de Sitter-Unruh modified-inertia proposal sets the MOND acceleration scale by the dark-energy density (a0 = cH_Lambda/Z = 9.36e-11 m/s^2), so a0 evolves as a0(z)/a0(0) = sqrt(rho_DE(z)/rho_DE(0)). The companion note showed that IF the lightest neutrino is a swampland-tower state, its mass is locked to a0(z) with no free amplitude, freezing by recombination so the CMB imprints a ~25-45% lighter neutrino than today's -- a right-signed reading of the DESI 'negative effective neutrino mass' anomaly. The open objection was DEGENERACY: a constant neutrino mass with GENERIC dynamical dark energy also pushes the inferred Sigma m_nu below the oscillation floor. This note identifies and quantifies the observable that breaks the degeneracy. A GROWING neutrino mass changes the free-streaming fraction f_nu(z), suppressing structure growth ONLY below the neutrino free-streaming scale k_fs(z) -- a SCALE-DEPENDENT step in the matter power spectrum. Dynamical dark energy changes the expansion H(z) and is SCALE-INDEPENDENT. The scale-dependence is therefore a separator no equation-of-state mirage can reproduce (a qualitative free-streaming-vs-background fork, not a numerical coincidence). A real CAMB calculation confirms the step (~0.65% at the normal-ordering floor, ~1.4% at Sigma=0.10 eV), and a focused Euclid Fisher gives a detectability of ~1.4 sigma at the floor rising to ~3.9 sigma at Sigma=0.10 eV -- i.e. the test's power is HOSTAGE to the absolute neutrino mass, which JUNO (first data 2026) and KATRIN pin independently this decade. The prediction is ROBUST to the DESI DR1->DR2 update (the locked mass ratio moves by <0.05: 0.56-0.71 -> 0.57-0.74). HONEST SCOPE: founded-not-derived (the absolute m_nu is free; the framework forces the RATIO and the scale E_dS=rho_DE^(1/4)=2.24 meV, not the mass); conditional on alpha~lambda and nu=tower; the CAMB step uses a two-constant-mass bracket and the Fisher rough single-tracer values (a publication forecast needs full Boltzmann m(z) + a multi-tracer Fisher); DIES if w->-1; NEUTRINO/DARK SECTOR ONLY -- NOT a theory of everything (charged-lepton and quark masses untouched and walled). Contribution: converts the companion paper's qualitative 'testable by a joint fit' into a specific, physical, dark-energy-unfakeable observable (the free-streaming step) with a computed detectability that sharpens with the absolute neutrino mass. Reproducible from the public repository.