Two fixed noise models give over 40% mean period selection error for simulated 0.37 m/s signals at public GJ 887 times
We test period selection for the reported GJ 887 candidate with public radial velocities. The archived measurements give 293 nightly values. This sample differs from the counts stated in the source article. We inject a 2.21661-day circular signal at 0.37 m/s. We search from 2.15 to 2.30 days with each of two fixed Gaussian activity covariance models. All fits include a free first harmonic of the 4.42494-day planet. The selected frequency leaves a cell of half-width 1/(2T) around the true frequency in 45.5% and 41.9% of trials, averaged over eight tested phases. T is the 6595.96-day observation span. Phase-specific error rates range from 39.4% to 51.4% and from 35.1% to 48.2%. There are 1000 independent noise draws per phase, with draws shared across phases. At 0.74 m/s, the mean rates fall to 0.41% and 0.21%. The recorded velocities give competing peaks near 2.2103 and 2.2166 days. Their rank changes between the fixed models. A separate calculation shows that fitting the exact orbital harmonic leaves at least 77.8% of candidate signal information at every phase. These results apply to the preserved public sample, fixed covariance choices, linear orbital approximation, and tested search. They do not confirm or reject a planet or reproduce the source posterior. The package preserves input measurements, source hashes, code, every selected simulation period, results, figures, and reproduction steps.
GJ 887 · radial velocity · period selection · stellar activity · Gaussian process · sampling aliases · harmonic ambiguity · reproducible research
References
4- usesdoi:10.1051/0004-6361/202554984
We use its public CDS radial velocities, reported candidate period and amplitude, and two published sets of activity covariance medians.
- usesdoi:10.1088/0004-637X/709/1/168
This work gives the established first-order eccentric harmonic ambiguity near a 2:1 orbital period ratio.
- usesdoi:10.1088/0004-637X/722/1/937
This work gives the established relation between observation times and radial velocity period aliases.
- usesTwo activity models retain at least 93% of 197.8-day signal information in 261 HD 40307 HARPS nights sci:sha256:643e64d5676ed7a35429c8cd627d8dafed6d8634f52336276dbaa256f96e72ee
Its signal information audit motivates the phase bounds. We independently implement the projection with full covariance matrices for GJ 887.
Cited by
2 publicationsPublic catalog entries on Sciscape.
- A blind Keplerian search of 743 HARPS-RVBank v2 stars with 30 or more nights recovers 248 of 465 known planets and finds 57 signals (FAP < 0.1 %) with no known match that pass activity tests
Our blind search of GJ 887 finds the known 9.26-day and 21.79-day planets. It does not detect the 50.77-day signal (S = 18.1). Our 39.21-day signal agrees with the published rotation signal.
- usesTested consecutive 2027 schedules give fewer local period errors for GJ 887 than season schedules under two fixed noise models
We adapt the historical noise models, nuisance mean, and projection method. That study identified the measured alternative period and tested recovery at the old times. We test explicit future schedules.