Computing Solar Eclipses — Research

Validation and error budget

workingupdated 2026-09-15validationerror-budgetsolar-radiusdelta-tpath-edge
  • Only edge observations test a prediction. A 0.37 arcsecond change of solar radius moves a 2017 central-line duration by 1.8 s and the duration a few hundred metres inside the southern limit by 19.3 s, so every quantitative test since 1973 has been made at or near a path limit 1.
  • Four independent campaigns since 2010 put the eclipse solar radius between 959.95 and 960.01 arcseconds, about 0.32 arcseconds above the 959.63 arcseconds nearly every product uses, which is 610 m per path limit 2 1 3 4.
  • The cleanest 2024 edge test observed 13.7 s of totality where six public products predicted 12.9 s to 65 s. Only a true-limb model with the larger radius came within a second 5.
  • At the edge the budget is dominated by three terms of 0.6 to 3 km each: the solar radius, the lunar limb if omitted, and terrain if omitted. Everything else is below 100 m and 0.5 s once ΔT is refreshed and a profile is used 1 6.
  • A five-level test protocol can be assembled from published cases, from the 1961 Explanatory Supplement worked examples through the 2001 Lusaka limb corrections to the Vale, Stephenville and bead-timing records 7 8 9.

What this topic covers

This topic asks how far eclipse predictions have actually been shown to be right. It collects the published comparisons of timed contacts and Baily's bead events against predictions, sets the major public predictors side by side and names the input behind each disagreement, and assigns seconds of contact time and metres of path edge to every term in the error of a modern prediction. It ends with reference cases and tolerances a new implementation can be tested against.

Notes in this topic:

  • Observed versus predicted: the IOTA and IOTA/ES edge campaigns from 1973, the 2017 Vale and Thermopolis results, the 2023 Cape Range results and the 2024 Stephenville experiment, with the measured offsets.
  • Predictor disagreements: the inputs the NASA SVS, Espenak, Jubier, Irwin, USNO, timeanddate, Occult, Stellarium and Photo Ephemeris products state, and how far their edges and durations differ.
  • Error budget and validation protocol: the sourced per-term budget, the record of ΔT model validation, and the five-level test protocol with tolerances.

What this topic changes for the pipeline

The solar radius stops being a constant and becomes a configuration item with a stated uncertainty and two documented modes. Every product carries a per-run error estimate and draws its path limits as a band rather than a line. The five-level protocol becomes an automated test suite, with each reference case's constants pinned beside its tolerance.

References

  1. 1peer-reviewed Quaglia, Irwin, Emmanouilidis & Pessi (2021), Estimation of the Eclipse Solar Radius by Flash Spectrum Video Analysis, ApJS 256:36 Full PDF read (var/downloads/quaglia2021_flash_spectrum_ApJS.txt). Flash-spectrum video from a site a few hundred metres inside the 2017 southern limit near Vale, Oregon. S = 959.95 +/- 0.05 arcsec. Gives sensitivity of duration and limit distance to solar radius, and compares Irwin's model with Occult and Solar Eclipse Maestro.
  2. 2peer-reviewed Lamy, Prado, Floyd et al. (2015), A Novel Technique for Measuring the Solar Radius from Eclipse Light Curves, Solar Physics 290, 2617 Abstract page read. 17 photometer determinations over 2010, 2012, 2013, 2015 with Kaguya limb data. Average 959.99 +/- 0.06 arcsec at 540 nm.
  3. 3trade Guhl (2023), Baily's Beads Observation during the Hybrid Solar Eclipse 2023 April 20, Journal for Occultation Astronomy 2023-4, pp. 12-15 Full PDF read (var/downloads/JOA2023_4.txt). Northern limit of the total segment, Cape Range, Western Australia. Sixteen bead timings, mean correction +0.38 arcsec, result 960.01 +/- 0.12 arcsec.
  4. 4trade Adalian (2024), April 8 solar eclipse maps are wrong along the edges, EarthSky Read via WebFetch summary. Journalism quoting C. Alex Young (NASA GSFC) and giving the 2,000 ft (610 m) edge shift for 959.95 versus 959.63 arcsec.
  5. 5trade Besselian Elements team, Experimentally Testing Eclipse Maps Accuracy (2024) Read via WebFetch summary. Stephenville, Texas, 2024 April 8. Observed totality 13.7 s (C2 18:39:06.6, C3 18:39:20.3 UTC) versus six predictions from 12.9 s (Irwin) to 65 s (timeanddate). Authors' own experiment, so trade grade.
  6. 6primary Espenak, Total Solar Eclipse of 2024 Apr 08, NASA GSFC interactive Google map page Full HTML read via curl (var/downloads/gsfc_SE2024Apr08Tgoogle.html). VSOP87/ELP2000-85 ephemerides, Delta T = 70.6 s, no limb profile, limits may shift 1 to 3 km, durations 1 to 3 s, greatest-duration point 10 to 20 km.
  7. 7peer-reviewed HMNAO and USNO (1961), Explanatory Supplement to the Astronomical Ephemeris, section 9 examples Local scan text read (var/downloads/es1961_djvu.txt). Worked examples 9.2 to 9.9 for the eclipse of 1961 February 15, usable as hand-checkable reference cases.
  8. 8primary Espenak & Anderson (2001), Total Solar Eclipse of 2001 June 21, NASA TP-2001-209484 Local text read (var/downloads/TP209484_2001.txt). DE200/LE200, Watts corrections of 0.4 arcsec, graze-zone accuracy +/- 0.3 arcsec, advice to stay 1 km inside the interior limit, Elev Fact terrain factor, consumer GPS +/- 100 m, worked Lusaka limb-correction example.
  9. 9trade Guhl & Tegtmeier (2018), Baily's Beads Observations during the Total Solar Eclipse 2017 August 21, Journal for Occultation Astronomy 2018-3, pp. 19-21 Full PDF read (var/downloads/JOA2018_3.txt). Two IOTA/ES stations, Thermopolis WY 700 m inside the northern limit and Cape Girardeau MO region. Bead-by-bead radius residuals reduced with Occult 4.5.3.0 using LOLA and Kaguya.

In this section

Observed versus predictedPublished comparisons of timed contacts and Baily's bead events with predictions, from the IOTA edge campaigns to the 2024 edge experiments, with the measured offsets in seconds and metres.workingPredictor disagreementsWhere the NASA SVS, Espenak, Jubier, Irwin, USNO, timeanddate, Occult, Stellarium and Photo Ephemeris products differ in inputs, and by how many metres and seconds at the path edge in 2017 and 2024.workingError budget and validation protocolThe contribution of each input to the error of a modern eclipse prediction, in seconds of contact time and metres of path edge, the record of ΔT model validation, and a test protocol with reference cases and tolerances for a new implementation.working