A Raw SPARC-to-KiDS Split-Kernel Transfer Test of Boundary-Gradient Lensing in the Emergent Condensate-Superfluid Medium Framework

This paper reports a raw SPARC-to-KiDS transfer test of the ECSM boundary-gradient lensing claim. A previously frozen SPARC V3 boundary-gradient source package, based on 175 disk galaxies and 3391 rotation-curve points, is transferred to an observed KiDS/2dFLenS weak-lensing bandpower vector using the observed P_E_gk signal, multipole coordinate ell, and the full 80 x 80 covariance matrix. The corrected ECSM optical closure is implemented as a split electric/magnetic response, with n^2 = Z_E/Z_B, while the common-scalar branch Z_E = Z_B is treated as a propagation null. The full split-kernel branch gives chi^2 = 134.37, compared with chi^2 = 145.79 for the no-boundary branch, chi^2 = 1486.25 for the common-scalar-only null, and chi^2 = 1105.80 for a single shuffled-boundary null. The full branch improves over no-boundary by Delta chi^2 = 11.42 for 80 data points, rejects the common-scalar propagation null by Delta chi^2 = 1351.87, and beats 1000 shuffled-boundary null realisations. A leave-one-bin jackknife gives 25 tests, with the full branch beating the no-boundary branch in 24/25 cases and the common-scalar branch in 25/25 cases. The final validation status is PASS WITH JACKKNIFE. One leave-one-source-bin test shows a small reversal, so the result is presented as a strong preliminary cross-domain validation of the ECSM split-kernel structure, not as proof of the full theory.