Eccentric terms remove at least 90% of local information near half periods at GJ 1002 times with known periods fixed

We test how free eccentricity changes information for an added circular velocity signal at the public GJ 1002 observing times. We use 139 times, internal errors, and instrument labels. We do not use measured velocities. The two known periods stay fixed at 10.3465 and 21.202 days. The primary errors are independent, with an added 1 m/s term in quadrature. We compare a circular model with its local eccentric extension. Generalized eigenvalues give bounds over every signal phase. At least 90% of local information is removed inside 5.171257 to 5.175069 days and 10.593085 to 10.609720 days. We also fit 384 noise-free additions with two exact Kepler orbits and eccentricities from 0 to 0.3. At both half periods, the 16 tested phases leave less than 0.9% of the circular squared distance at 0.1 m/s. They leave less than 7.3% at 0.3 m/s. Period shifts of 0.25% leave much larger distances. The first-order ambiguity is established prior work. Our contribution is the schedule-specific interval map, finite orbit alternatives, and reusable code. The result depends on fixed periods, a circular baseline, and the stated errors. It is not a planet detection or a recovery probability. The package preserves the source data, provenance, fit parameters, figures, and independent numerical checks.

GJ 1002 · radial velocity · orbital eccentricity · half period · signal information · observing times · Kepler orbit · reproducible astronomy