A fixed 10.03 hour term reduces block prediction error near ZTF J0328-1219 in TESS Sectors 108 and 109 with selected spline smoothing

A selected 10.0324-hour term improves block predictions of TESS aperture flux near the known debris-transit white dwarf ZTF J0328-1219. We test it with a fixed 11.3604-hour main pattern whose Fourier coefficients can have smooth time envelopes. Five outer folds hold out whole time blocks. Inner folds choose spline spacing and smoothing from training target values alone. In the primary four-pixel, half-day-block test, adding the shorter-period term reduces mean squared error from 1.65097 to 1.51769 (electron/s) squared in Sector 108, a gain of 8.07%. The partial Sector 109 record through 2026 October 18 gives a gain of 11.66%. All 36 tested aperture, block, origin, and harmonic settings per sector give positive gains. Separate least-squares calculations check the primary predictions and selected inner scores. Unregularized spline failures and the adaptive smoothing amendment are preserved. The periods come from prior SCI fits to some of the same photons. The results are not an independent detection or a calibrated significance test. A changing main pattern can represent a nearby frequency, so the gain does not establish a second orbit or a planet. The package includes exact public input images, provenance, code, all model scores, predictions, injections, figures, and offline reproduction instructions.

ZTF J0328-1219 · TESS · white dwarf · planetary debris · time series · blocked prediction · spline smoothing · changing waveform · reproducible research