UAP AnalysisIndependent · the declassified record

Observation taxonomy

normalized vocabulary across the corpus

Raw documents describe objects in inconsistent language. We map those descriptions onto a normalized vocabulary so they can be counted and compared — while preserving each original phrasing on the record page.

Reported shapes
disc
50
orb
37
ellipsoid
18
cylinder
15
irregular
10
triangular
4
football
2
Most-reported movements
high-speed76
hovering32
90-degree turn11
evasive7
launching sub-objects6
formation flight6
no sound6
instantaneous flash5
vertical ascent4
high-altitude3
directional flight northward3
circling3
stationary3
no trail3
tumbling3
rotating3
Sensor modalities
human eye156
visual-band (eye/EO)56
infrared / FLIR35
radar15
RF / ELINT1

These are physical sensing modalities(the human eye, visible-band/EO, infrared/FLIR, radar, RF ELINT) — kept distinct from the collection discipline (HUMINT, IMINT, SIGINT), which is a different axis. “Visual-band” covers both naked-eye and visible-band cameras; the source documents rarely separate them, so we don’t assert which.

Multi-source corroboration

202 cases carry a sensor tag

Events recorded by more than one independent sensing channelcarry more weight than single-source reports. We count channels conservatively — the human eye and a visible-band camera share one band, so they don’t count as two. “Cross-class” is the strongest tier: an optical channel (eye/EO/IR) and an independent radar or RF detection — confirmation by different physics.

Important caveat — corroboration here is as described by the source document, and provenance varies. Of the 13 cross-class cases, 7 come from primary records (debriefs, cables, incident reports) and 6 are famous cases summarised in a secondary study (marked below) — real events, but not primary multi-sensor data we hold. Treat the primary set as the stronger evidence.

Single-source
165
Multi-source
24
Cross-class corroborated
13
The 13 cross-class corroborated events