DOW-UAP-PR102 and PR103 are rare modern UAP cases where the public can watch the government-released infrared footage itself. One target leaves the frame. The other sits inside an auto-track box for more than a minute. What the release does not give us is just as important: no public range, speed, altitude, platform identity or confirmed explanation.
HERO MEDIA: authentic DOW-UAP-PR103 footage hosted by DVIDS / All-domain Anomaly Resolution Office. U.S. Government public-release media. The background is the official video player, not a reconstruction or AI image.
The United States Indo-Pacific Command submitted two infrared video records to AARO from a U.S. military platform operating over the East China Sea in 2024. The government released them separately as DOW-UAP-PR102 and DOW-UAP-PR103.
That provenance matters. These are not anonymous social-media clips and they are not reconstructed imagery. The footage is hosted through official U.S. military channels, with AARO named as the source organisation and the East China Sea identified as the location.
But official provenance is not the same thing as an exotic interpretation. A real military sensor can record a distant aircraft, balloon, surface object, atmospheric contrast or sensor artefact. The public evidence establishes a genuine military record of something not identified in the released material; it does not establish what that something was.
AARO's own release language warns readers not to treat its video description as an analytical judgment or factual determination about the event's nature or significance. The word unresolved means the public record does not supply a completed identification. It does not automatically mean anomalous technology, extraordinary flight or non-human origin.
For roughly the first twenty seconds, the infrared sensor keeps an area of contrast generally centred. The target then exits the top of the field of view. During the closing seconds, the sensor zooms out and in but the official description says there is no content visible.
What that establishes: a trackable contrast existed for part of the clip. What it does not establish: why it left the frame, its range, its true motion or its identity.
AARO describes the sensor panning to keep an area of contrast generally near the centre throughout the clip. From about twelve seconds onward, an auto-tracking reticle surrounds it and remains engaged for the rest of the released sequence.
What that establishes: the sensor could maintain a stable track on a contrast feature. What it does not establish: a physical shape, extraordinary acceleration, size or origin.
The strongest way to read this case is side by side. The public record gives PR102 and PR103 separate identifiers, separate clips and different tracking behaviour. Nothing released by AARO establishes that they are the same object, the same sortie or even the same encounter.
| Question | PR102 | PR103 |
|---|---|---|
| Official title | DOW-UAP-PR102 — Unresolved UAP Report, East China Sea, 2024 | DOW-UAP-PR103 — Unresolved UAP Report, East China Sea, 2024 |
| Released duration | 36 seconds | 1 minute 16 seconds |
| Sensor description | Infrared sensor aboard a U.S. military platform | Infrared sensor aboard a U.S. military platform |
| What the clip does | The sensor tracks an area of contrast for roughly the first 20 seconds; it then exits the top of frame. The sensor subsequently zooms with no content visible. | The sensor pans to keep an area of contrast generally near centre; from about 00:12 onward an auto-tracking reticle surrounds it for the remainder of the clip. |
| What is directly established | A trackable infrared contrast was present for part of the clip and was then lost from view. | A persistent infrared contrast was stable enough for the tracking system to maintain auto-track for about a minute. |
| What is not established | Range, size, speed, altitude, extraordinary manoeuvre, physical shape or identity. | Range, size, speed, altitude, extraordinary manoeuvre, physical shape or identity. |
| Same event? | Not established by the public release. Shared year, region and submitting command are not enough to merge the records. | |
A combined retelling can quietly create stronger evidence than the government actually released. PR102's loss-of-track behaviour cannot be used to describe PR103, and PR103's sustained auto-track cannot be back-applied to PR102. Treating them separately preserves what is genuinely known and exposes exactly where the public record ends.
AARO's wording is unusually important here. The office does not describe a disc, sphere, drone, aircraft or vehicle. It describes an area of contrast.
Infrared systems display differences in thermal radiation rather than a natural-colour view of the scene. A bright or dark patch can represent a real warm object, a cold object, exhaust, cloud structure, reflected radiation, a distant target blurred by the optics, or other contrast that the sensor can lock onto without revealing a clean physical silhouette.
An auto-track box is therefore evidence that the tracking algorithm found a sufficiently stable feature to follow. It is not, on its own, evidence that the feature was a solid craft or that its apparent movement across the display equals its true movement through three-dimensional space.
That distinction is where some retellings of PR103 run ahead of the primary source. Claims that PR103 shows “multiple UAP expanding and contracting,” for example, go beyond the primary wording. AARO describes an “area of contrast” held by the sensor and auto-tracker; its public caption does not identify multiple craft, expansion/contraction as physical behaviour, or any anomalous performance.
A Jan. 1 date appears in DVIDS metadata for both records. Because the official narrative describes the incident only as occurring in 2024, this page does not treat that catalogue date as proof that both clips were recorded during the same event.
A distant aircraft, balloon, drone, surface vessel, atmospheric feature or sensor/optical effect can all generate an infrared contrast that is easier for a tracker to follow than for a viewer to identify. Platform motion and parallax can also make a relatively ordinary target appear to move unexpectedly when range is unknown.
The released footage is too information-poor to eliminate those categories. PR102's loss of the target and subsequent zooming on empty sky can be read as a sensor/operator reacquisition problem without requiring any extraordinary manoeuvre.
The important counterpoint is equally simple: AARO and USINDOPACOM did not publish an identification. PR103's stable auto-track demonstrates that the sensor was following something persistent enough to hold for more than a minute, and the government still released the record under an unresolved UAP designation.
If the underlying classified or unreleased data contains range, platform motion or other sensor information that defeats mundane explanations, the public clips do not disclose it. The responsible position is therefore not to invent certainty in either direction.
Raw or calibrated sensor metadata; exact platform motion; line-of-sight angles; range estimates; contemporaneous radar or other sensor records; operator testimony; weather data; and a clear statement on whether PR102 and PR103 came from the same sortie would transform what can be tested. Until then, the mystery sits in the gap between an authentic military track and an incomplete public dataset.
They are separately numbered U.S. government infrared video records submitted by U.S. Indo-Pacific Command to AARO and labelled as unresolved UAP reports over the East China Sea in 2024.
The public record does not establish that. They share the same broad location, year and submitting command, but AARO/DVIDS released them as separate records with different durations and tracking sequences. Until a common sortie, time, platform or case identifier is released, the safest reading is to keep them separate.
It proves the infrared tracking system could maintain a lock on an area of contrast for most of the released clip. It does not by itself prove the target was a solid craft, determine its range or establish extraordinary performance.
No responsible speed can be derived from the public video alone because the release does not provide the target's range, full line-of-sight geometry or calibrated kinematic data.
Yes, those remain plausible categories on the public evidence. The official release does not identify the target, and the available footage does not contain enough public data to exclude ordinary distant objects or sensor/parallax effects decisively.
Because the media is genuine, military-sourced and openly viewable, while the identity remains undisclosed. It is a useful example of the difference between authentic UAP evidence and evidence of extraordinary technology.
Editorial note: competitor pages were used only to identify search-intent and coverage gaps. Claims here are written independently and checked against the official records above.