Evidence
fact ⇄ what kind of claim · live
Cluster physics: 600 ft/min up, 800-1,000 ft/min down, a 46-60 kft shelf
NYU flight data via the 1995 USAF report · rates used as written
MEASURED
The wind stack: NE below, NW at the tropopause, W above
Moore's method (Orogrande pibal + Flight 5) · magnitudes tuned to endpoints
MODELLED
Flight 5, June 5: 0516 MST, 5 h 50 min, down near Roswell, 12 intact
tracked end to end · the control case this bench is pinned to
MEASURED
Flight 4's path past Arabela is Moore's reconstruction
published 1995-97 · arithmetic contested by Rudiak & Sparks
MODELLED
Crary's diary holds both sentences: “no balloon flights” and “flew … cluster”
4 June 1947 · each camp reads a different sentence
REPORTED
Brazel's inventory: rubber, foil, paper, sticks, tape, eyelets, ~5 lb
Roswell Daily Record, 9 Jul 1947 · his own words
REPORTED
The flowered tape: radar targets built by a toy factory
Moore and colleagues' recollection · the file's cleanest before/after
REPORTED
Whether the ranch debris was Flight 4 is a reading
the bench computes the drift; the file stays open
READING
View
The mapthe drift, integrated
The train600 ft, to scale
The debrislist against list
The file1947-1997
Flight
Flight 4Jun 4 · the contested one
Flight 5Jun 5 · the control
One balloonwhat Ramey showed
Customyour train, same winds
…
Playback & layers
Playbackhours compressed to seconds
Rate
300× 600× 1800×
Tracked vs dashedthe seam at Arabela, drawn
Winds-aloft columnthe three storeys, live
Terrainridges and White Sands, MODELLED
The retellingsdebris view: what 1978+ added
Readout · live from the integrator
Flight in force
…
Ceiling reached
—
Time at float
—
Total flight
—
Down at
—
Balloons at landing
—
Range
—
To the Foster Ranch
—
…
What is measured vs modelled here
Measured: the train's bill of materials, the climb and descent rates, Flight 5's clock end to end. Reported: Crary's diary, Brazel's inventory, the toy-factory tape, every quotation in the file. Modelled: the wind tables (Moore's layer structure, magnitudes tuned to the recorded endpoints), the terrain drawing, Flight 4's entire path past Arabela. Reading: whether the ranch debris was Flight 4. The bench computes the drift and stops there.
The drift model
Vertical: a cluster climbs at the recorded ~600 ft/min while lift beats weight, floats at its ceiling, loses one balloon per attrition tick after sunrise, and sinks at the recorded 800-1,000 ft/min (gentler below 25,000 ft) once lift fails. Horizontal: dx/dt = W(h), a layered winds-aloft table integrated along the altitude profile — northeast through the troposphere, northwest in the transition layer above the tropopause, west above. The tables reproduce Flight 5's recorded endpoint (5 h 50 min, near Roswell, 12 balloons intact) and Moore's published Flight 4 endpoints, and wear a MODELLED tag because that is what they are.
Where this drawing takes liberties
The terrain is a sketch; the debris-field ellipse is drawn oversized to be clickable; the wind magnitudes are tuned, not archival, because the archival pibal sheets are not published in machine-readable form; the attrition is a rate, not a per-balloon history. The train's parts list, the recorded rates, the diary sentences and every quotation take no liberties.
▶ Story · 15 chapters
⟳ Reset