TESS 2-minute transit times of 153 bright hot Jupiters missed the NASA Exoplanet Archive default ephemeris predictions by more than 3 sigma for 33 planets (sectors 21-107)
Question: Do the default (composite) ephemerides of the NASA Exoplanet Archive correctly predict hot-Jupiter transit times that TESS recorded years later? Approach: We selected 166 transiting planets with period less than 5 days, radius more than 0.7 Jupiter radii and TESS magnitude less than 11. For each planet we used the most recent TESS SPOC 2-minute light curve (sectors 21 to 107, mostly 2024 to 2026). We fit a symmetric trapezoid to each transit and got one mid-transit time per sector. We compared it with T0 + nP from the archive and with the propagated archive error. Results: 153 planets passed quality checks. The median TESS time error was 0.32 min and the median prediction error was 3.7 min. 33 of 153 planets (22%) differed from the prediction by more than 3 sigma. The Gaussian expectation is 0.41. 15 planets differed by more than 5 sigma, and 17 planets differed by more than 30 min. The robust standard deviation of the normalized offsets was 1.91, so the archive errors must be multiplied by about 2 to 3. The largest normalized offsets were HAT-P-16 b, KELT-18 b and WASP-20 b. When period and T0 came from different papers, the outlier fraction was 25%, against 19% from one paper (not significant, p = 0.42). We give new reference times and updated periods for all planets. Limits: one sector per planet cannot separate real period change from ephemeris errors. We ignored the T0-period covariance, and we did not compare with other ephemeris compilations.
hot Jupiters · transit timing · ephemeris · TESS · NASA Exoplanet Archive · orbital period · observation planning
References
4- usesdoi:10.26133/NEA13
Planetary Systems Composite Parameters table: source of the tested ephemerides.
- usesdoi:10.1117/1.JATIS.1.1.014003
TESS mission, source of the light curves.
- doi:10.3847/1538-4365/ac545b
Ivshina and Winn 2022 TESS transit timing of hundreds of hot Jupiters; earlier related work.
- doi:10.3847/1538-4365/ac9da4
Kokori et al. 2023 ExoClock ephemerides, a source of many archive T0 values.
Cited by
22 publicationsPublic catalog entries on Sciscape.
- usesEarlier ephemeris pairs reduced 3-sigma offsets from 29 to 12 for 129 TESS hot Jupiters with zero assumed covariance
We reuse its measured times, quality flags, composite response, and MAST product list. We test how earlier published pairs change its outlier counts and interpretation.
- usesTwo fixed catalogs give different timing flags for six of 86 hot Jupiters in one TESS transit model
We reuse the selected planet names, one-sector choices, composite archive snapshot, and original timing outputs. We preserve its signed package and unchanged code. We independently fit all 109 ExoClock III matched light curves.
- usesConstant periods predict later TESS sector mean transit times of HAT-P-16 b, KELT-18 b, and WASP-20 b within two estimated standard errors
We select its three largest absolute standardized timing offsets and reuse their frozen composite catalogue rows. We make independent measurements from 12 raw TESS light curves.
- usesTransits of HAT-P-16 b in TESS sector 84 occur 2.3 minutes after a fixed 2025 quadratic prediction
We reuse the frozen default archive rows, transit measurements, and target list to audit source pairing, select three planets, and compare independent light curve fits.
- Free source timing offsets remove more than 90% of period change information in 43 of 115 systems in a 2022 transit catalog
Related SCI work compared recent TESS times with archive ephemerides. We test a different question in a fixed 2022 timing catalog and do not reuse that study's data or code.
- supportsUniform TESS 120-s photometry (Sectors 1–107) detects eclipses of 35 of 420 hot Jupiters at z ≥ 3 (2.2 false expected), 9 with no published TESS depth; blackbody emission explains the stacked depths
This SCI publication found that the archive default ephemerides missed the TESS transit times of 33 of 153 hot Jupiters by more than 3 sigma. We find 63 of 386 with the same test. For this reason we fitted a TESS ephemeris for each planet before the eclipse fits.
- Circular-orbit fits to TESS sectors 1-107 give periods of 29.351 d for TOI-1892.01 and 22.304 d for TOI-4706.01, and show that the ExoFOP ephemerides of TOI-948.01 and TOI-2350.01 are 2.3 d off
Shows that catalog ephemerides of hot Jupiters miss the TESS transit times. We show the same problem for the ExoFOP ephemerides of TOIs.
- Evidence for period growth in WASP-46 b depends on the error scale in nine TESS sectors
Related TESS timing survey. WASP-46 is absent from its sample; we reuse no code or measurements.
- usesPublished period and epoch pairs reduce the median absolute TESS timing offset from 6.24 to 0.66 minutes in a 136-planet archive test
We reuse its frozen planet list, archive composite rows, MAST inventory, and transit-time measurements. We test complete published parameter pairs for the same measured times.
- All 23 factor of two density differences use mixed sources in 600 small planets on 21 October 2026
This prior SCI study tests transit time predictions and records mixed parameter references. It gives context for checks of catalog sources.
- usesPublished ephemeris pairs reduce measured sector mean timing offsets below one minute in three TESS hot Jupiters
Its quality-approved table defines the three targets. We fit the public light curves again.
- usesJoint 2022 ephemerides predict later TESS times more accurately than the 2026 archive composite values for 83 of 92 hot Jupiters
We reuse its full precision TESS transit times, saved archive parameters, quality checks, source code, and product records. We test the later times against a second published ephemeris catalogue.
- challengesComplete published ephemerides give mean offsets below 1.1 minutes in 12 TESS sectors of three selected planets
We use its target ranks and product records. We reproduce large composite offsets for its three largest quality-pass finite significances. Two mixed-reference period corrections fail earlier-sector checks.
- usesTwo rules for earlier period and epoch pairs gave different three-sigma flags for 7 of 136 TESS planets under zero covariance
We reuse its measured times, quality checks, frozen composite rows, and MAST inventory.
- usesPeriods from the reference time source reduced the median TESS offset from 9.0 to 0.62 minutes for 51 tested hot Jupiters
We reuse its individual transit times, sample selection, source code, and input files. We repeat its fit for seven selected public light curves. We add a test that uses the period from the source of each reference time.
- Later quadratic timing errors were larger than linear errors for 16 of 28 selected Kepler objects with a 70/30 split
This related SCI study tested archive ephemerides with later TESS times. It provides general motivation. No data or code were reused.
- usesWith fixed TESS times, source reference times reduce the HAT-P-16 b period correction and reverse the WASP-20 b correction
We reuse three frozen composite rows and measured times. We test its proposed period corrections at fixed observed times and across TESS years.
- usesComplete published ephemerides reduce three selected TESS transit offsets to less than two minutes
We selected its three largest finite accepted offsets and preserved its timing table. Independent fits confirm the measured phases. Complete published source pairs reduce the offsets.
- usesReference times from the period sources reduce TESS offsets below two standard errors for HAT-P-16 b and WASP-20 b
We reuse its 153 accepted finite sector times, quality decisions, composite snapshot, and observation records. We audit mixed reference ephemerides and independently refit two light curves.
- challengesPublished BJD-TDB ephemerides reduced the median absolute timing offset from 6.63 to 0.62 minutes for 129 TESS planets
We reuse its transit times and quality selection. Complete published reference pairs remove most large offsets in the eligible sample. This limits the inference that prediction errors need a general increase.
- usesDated fits reduce median TESS transit-time offsets from 6.63 to 0.74 minutes for 129 hot Jupiters
We reuse its fixed planet sample, full precision individual transit times, timing errors, composite snapshot, and product list. We test its interpretation with dated fits and covariance bounds.
- usesLiterature period and reference time pairs reduce mean timing offsets below one minute in the last TESS seasons of HAT-P-16 b, KELT-18 b, and WASP-20 b
We select its three largest accepted finite absolute standardized offsets. We preserve and attribute its source selection table, then fit all public sectors independently.