JAL 1628: The Star Bench

The Night view is 3D because the subject is a piece of celestial geometry lived from a cockpit: where the lights stood relative to the nose, minute by minute, through every heading change. The 747 flies the tape's own clock over moonlit interior Alaska; the star field overhead is computed for that exact evening and mounted on an equatorial frame driven by the same arithmetic the tune script audits; the amber lights are drawn at the record's clock calls and labelled as testimony. Press Run and 17:05 to 17:57 plays in under two minutes: the first lights, the arrays, the pale band, the silhouette, the 360, United 69.

The Bearings view is the case's core ledger: the record's own bearing calls as amber dots, from '30 degrees left front' to 'seven, ah, eight o'clock', plotted against the computed clock-position and elevation curves of Jupiter, Mars and the Moon. The reported drift around the clock face happens while the aircraft's heading walks from 215 through 255 and back to 198: a fixed patch of sky slides backward around the nose exactly this way.

The Radar view prices each scope on its own lane: the crew's eyes, the weather radar's single green echo, Anchorage Center's few primaries, the military's surge return and 'flight size of two', and the aircraft sent to look. Filled marks found something; hollow marks looked and found nothing. The FAA's split-return finding and Maccabee's objections ride the same lane, side by side.

The File view is the claims on ten cards: the size that grew tenfold in six weeks, the ninety-fold distance disagreement, the three-way witness split, the camera that photographed nothing, the FAA conclusion, the green return, Callahan, the captain's own January retraction, the career cost, and both auditors checked against the sky. It closes on two verdicts at equal size.

This station's wiki entry gives the case one paragraph: the captain, the mothership, Callahan, the meeting that 'never happened'. A paragraph cannot hold what makes this file genuinely interesting: it is that rare radar-visual case where nearly every minute is on tape. The Anchorage Center recording, the FAA tracking data, three separate crew interviews, and two duelling auditors, Klass and Maccabee, reconstructing the same evening, are all publicly readable. Where there is tape, there can be pricing; where there can be pricing, this station builds a bench.

It also picks up two older threads. INST-54, the Belgian Wave, built this station's exact-ephemeris pipeline for 'the sky that night'; this bench reuses it whole and adds the Moon with topocentric parallax. When Terauchi declined a military intercept, the lesson he says he remembered is the one INST-36, Mantell 1948, prices: do not send fighters after a light you have not identified. And the mechanism Klass offered for the first act, moonlit ice-crystal cloud, shares its atmospheric optics with INST-41, the Fata Morgana bench. Read side by side, the sentence 'a captain reported a light' weighs what it honestly weighs.

One exact sky, one exact geometry, zero fitted constants. The panel tags every number exact, modelled, reported or reading.

az/el = f(GMST(JD), α, δ, φ(t), λ(t)) · rel = az − hdg(t) · R = v²/(g·tanφ) · S = d·θ

The sky, audited. GMST from the IAU 1982 expression; the J2000 catalogue rotated to the aircraft's own interpolated position; Jupiter and Mars from Standish mean elements with the standard magnitude laws; the Moon from a truncated Meeus series with topocentric parallax. Every curve the instrument draws is asserted against JPL Horizons values in the tune script before it is allowed on screen.

The geometry, solved. Headings come from the reported route fixes by great-circle bearing, and land on the tape's own 215/255/240 sequence without being typed. Relative bearing is azimuth minus heading; a clock hour is thirty degrees; the horizon dip at FL350 is 3.3 degrees, a number one published rebuttal got wrong by half.

The manoeuvres, priced. The 360 is R = v²/(g·tanφ): the stated 30-degree bank at 500 knots buys 6.3 NM and 4.8 minutes, and the tape agrees to seconds. The one manoeuvre the case records in detail is ordinary flight mechanics.

The sizes, honest. S = d × θ, with the assumed distance as a dial because the crew's own numbers span 500 feet to eight miles. Jupiter's true disc that evening was 43 arcseconds: a pure point, unrangeable by eye, which is what lets a planet fly formation forever.

What is deliberately not computed. No identification of the first-act arrays (the mechanism on offer is carried as a candidate with its objections), no verdict on what three witnesses saw, no psychology. Where the record is silent or contested, the bench carries a label instead of a guess.

The sky, the bearings, the turn and the size arithmetic are exact and machine-audited; the staging of the lights and the terrain is modelled and labelled; the chronology, the quotes, the radar events and the aftermath are reported as documented; and what the unexplained first act amounts to is a reading, for which the bench hangs two verdicts at equal size and adds no third.

A clock, a sceptic switch and one distance dial: between them they hold the whole public argument about JAL 1628.

• The clock, with Run. 17:05 to 17:57 at thirty times real speed, with the record's fixed moments marked: 17:09 contact, 17:17 the arrays, 17:23 the pale light, 17:30 the silhouette, 17:37 the 360, 17:50 United 69. Scrub anywhere; every view shares the same clock.
• The sceptical annotations. Off, the evening plays as the crew reported it. On, the planets wear halos, each phase wears the reading the later analyses gave it, and the radar lanes carry the FAA's own words.
• The assumed distance. The single number every giant in this file is built on, and the one number nobody measured. Slide it from the captain's written 500 feet to the radar's eight miles and watch the claimed object grow from a DC-8 to a hillside of light while the testimony stands still.

Six claims that orbit JAL 1628, from Klass's identification to the reading the case still carries, with who made each and where it lands once the sky is recomputed.

1. Press Run and fly the whole evening once. Watch the amber lights and the gold planet at the same time. Notice which one does something a sky cannot do, and which one never does.
2. At 17:37, watch the 360 from above. Zoom out. The circle is 6.3 miles wide and takes 4.8 minutes, exactly as a 30-degree bank at 500 knots must. Mid-turn the light disappears; out of the turn it is back where a fixed sky would keep it.
3. In the Bearings view, follow the amber dots. They start at 11 o'clock, where no planet stands, and end at 7-8 o'clock after two right turns. Now follow Jupiter's curve through the same minutes and decide how much of the drift belongs to the light and how much to the nose.
4. Slide the distance dial from 500 ft to 8 NM. The claimed object passes from DC-8 to double-carrier without a single word of testimony changing. That factor of ninety is the entire size case.
5. In the Radar view, toggle the annotations and count filled versus hollow marks. Then click the controller's last entry: 'a few primary hits... never a good track, ever'. That sentence was spoken on the night, not written by a sceptic.
6. End on the File's two verdicts. They do not fight. The pacing was the sky's; the arrays were never identified. Holding both at once is what this file actually supports, and it is rarer than either alone.

The honest line between what is exact, what is modelled here, what the record reports, and what is a reading:

In one line: the sky over the route (audited against JPL Horizons), the bearing geometry, the turn physics and the size arithmetic are EXACT with no fitted constant anywhere; the staging of the reported lights, the terrain and the interpolated path are MODELLED and labelled; the tape chronology, the crew interviews, the radar events, the FAA finding, the sceptical case, its rebuttals and the aftermath are REPORTED as documented; and what the unexplained first act amounts to is a READING, for which this bench hangs two verdicts at equal size and adds no third. Fly the evening, check the sky yourself, and decide.

• Bruce Maccabee, "The Fantastic Flight of JAL1628", International UFO Reporter, March/April 1987: the minute-by-minute chronology from the AARTCC tape (17:09:20 initial contact to the 18:20 landing), the crew interview excerpts (Terauchi, Tamefuji, Tsukuba), the tracking-data headings (magnetic 182/222, true 215/255), the descent to FL310, the 360-degree turn timing, the weather-radar sketches, the camera episode, and his objections to the CSICOP release. The bench's spine chronology follows this reconstruction.
• Philip J. Klass, "FAA Data Sheds New Light on JAL Pilot's UFO Report", Skeptical Inquirer 11(4), Summer 1987: Jupiter at magnitude -2.6 near azimuth 143, Mars about 19 degrees right, the nearly full moon almost directly behind the flight, the sun about 11 degrees below the horizon, the moonlit ice-crystal mechanism for the first act, the green-return argument, the United 69 geometry, and the crew-interview quotes; with the CSICOP press release of 22 January 1987 (Sanduleak and Stephenson's computations) as its earlier draft.
• The FAA release of 5 March 1987 (spokesman Paul Steucke, Alaskan Region): the split-radar-return finding from the FAA Technical Center (specialist Dennis Simantel's uncorrelated primary-beacon attribution), the statement that the agency was unable to confirm the event, and the on-the-record note that the FAA did not analyse the visual sighting. The complete ~377-page FAA package (chronology, statements by controller Carl Henley and inspectors Wright and Mickle, transcripts, drawings) was sold publicly and is scanned at nicap.org and archive.org.
• Robert Sheaffer, "JAL 1628: Capt. Terauchi's Marvelous 'Spaceship'", Skeptical Inquirer 38(6), Nov/Dec 2014: the Callahan account and its contradictions (Ron Pandolfi's and Maccabee's recollections of the meeting), the January 11, 1987 second sighting and its agreed village-lights explanation, the pipeline-pumping-station episode, and the size-claim escalation through the press.
• John Callahan's Disclosure Project testimony (2001) as the claim under audit: the home-video radar playback, the briefing to Administrator Engen, and the 'this meeting never happened' account; carried beside the attendees' recollections above.
• JPL Horizons (ssd.jpl.nasa.gov), queried for this build: topocentric azimuth, elevation and magnitude of Jupiter, Mars and the Moon from the Fort Yukon leg for 02:10-03:00 UTC on 18 November 1986. The bench's own Standish-elements and Meeus-series ephemeris is asserted against these values, to 0.35 degrees for the planets and 0.25 for the Moon, in scripts/jal-1628-tune.mjs.
• Astronomy and flight mechanics: GMST per the IAU 1982 expression (Meeus, Astronomical Algorithms, which also supplies the truncated lunar series); Standish, "Keplerian Elements for Approximate Positions of the Major Planets" (JPL, 1800-2050 table); the Astronomical Almanac magnitude laws for Jupiter and Mars; the level-turn relation R = v²/(g·tanφ); the 4/3-earth horizon-dip and radar-horizon expressions. All constants are solved and printed by the re-runnable tune script; this bench fits no constant anywhere.
• Aircraft model: "Japan Air Lines Boeing 747SR-46 (JA8119)" (https://skfb.ly/pKPPs) by LB&SCR E2, licensed under Creative Commons Attribution 4.0 (http://creativecommons.org/licenses/by/4.0/). Modified for this bench: landing gear and wheels removed at the node level for the in-flight configuration. The SR passenger variant stands in for the case's 747-246F freighter: the same short-hump classic airframe, carrying a painted window band the freighter did not have.
• Cross-referenced instruments on this site: INST-54, the Belgian Wave, whose exact-sky pipeline this bench reuses; INST-36, Mantell 1948, the precedent Terauchi cited in declining an intercept; INST-41, Fata Morgana, for the atmospheric-optics family the ice-crystal mechanism belongs to; and the station's wiki entry on JAL flight 1628.