Kepler Data Validation Documentation

Within the Kepler pipeline, stars that have been identified with at least one Threshold-Crossing Event (TCE)—which is a series of at least three transit-like signals with a consistent period and sufficient signal-to-noise ratio—are put through a process called Data Validation (DV) (Wu et al. 2010). DV produces several data products that have been made available to the community:

Time Series Files

The DV time series files are available for each TCE in Q1-16, Q1-17 DR 24 and Q1-17 DR 25 pipeline runs. The format and contents depend on the run and are described below.

The collection of DV time series files can be downloaded from the Bulk Data Download page. Files for individual objects can be plotted in the time series viewer. Links to these plots are available by clicking on the information icon () next to an object in the TCE, KOI and Kepler Stellar interactive tables.

Q1-17 DR 25

The DR 25 DV time series files are available as FITS or IPAC ASCII tables files. The files contain the detrended, quarter-stitched light curves, the whitened light curve and the transit models. Details for the file contents are given in this document (KSCI-19079).

Q1-16 and Q1-17 DR 24

Each TCE is given in a separate ASCII file using IPAC table format. The KIC number, quarters processed and TCE number are included in the file name. The header includes this information as well as the midpoint of the first and last cadence in MJD (LC_START and LC_STOP), the total number of planets for this KIC, and the orbital period and transit epoch for this TCE from the TCE table.

The data columns in the file are:

  1. TIME (in BKJD = BJD – 2454833)
  2. CADENCE_NUMBER
  3. INIT_FLUX_PL(e-/e-)
    • Unfilled, normalized, harmonic-removed input to planet search algorithm for this TCE. Previously identified TCEs have been gapped.
  4. INIT_FLUX_PL_ERR (e-/e-)
    • Uncertainty in the initial flux
  5. MODEL_LC_PL (e-/e-)
    • Filled, normalized model light curve calculated from the planetary fit to the TCE.
  6. RESIDUAL_FLUX (e-/e-)
    • INIT_FLUX_LAST_PLANET with model transits gapped – median(INIT_FLUX_LAST_PLANET with model transits gapped)
  7. RESIDUAL_FLUX_ERR (e-/e-)
    • Uncertainty in the residual flux

The light curves provided in the DV time series files do not exactly match those plotted in the PDF summary files. The flux curves plotted in the PDF summary files have additional processing, specifically, a median filter to remove long-period systematics (all plots except D) and a whitening filter (plot D).

If the DV process fails to converge for a given TCE, the MODEL_LC_PL flux will be given as NaN and the PDF summary file will say "DV fit results are unavailable." This happens only occasionally and occurs in systems with large systematic noise and for very deep transits.



Last updated: 22 September 2016