Unknown

The Three Lights That Held Formation Over the Indo-Pacific (DOW-UAP-PR47, 2023)

Japan (per the Department of War PURSUE Release 01 manifest; the DVIDS record itself publishes no location) · 2023 · Government file · Japan

A frame from the infrared footage released by the Department of War as DOW-UAP-PR47, showing the monochrome thermal scene in which the sensor holds three distinct areas of contrast near the centre of the display. The black bars around the edge of the frame are the Department's own redaction blocks, which cover the sensor telemetry and data burn-in throughout the clip. This is real released government footage, not a reconstruction or an illustration. (US Department of War / All-domain Anomaly Resolution Office, distributed through DVIDS. Public domain.)

For one minute and fifty-nine seconds a United States military infrared sensor held three separate points of heat in the centre of its frame, and the Department of War's own analysts recorded the only detail in the file that can be tested: the three never moved relative to one another.

What did witnesses see at Japan (per the Department of War PURSUE Release 01 manifest; the DVIDS record itself publishes no location)?

The record at the centre of this case is one video file, one minute and 59 seconds long, published by the United States Department of War as DOW-UAP-PR47, "Unresolved UAP Report, INDOPACOM, 2023". It went up on the Defense Visual Information Distribution Service on 8 May 2026 at 07:50, credited as a courtesy video from the All-domain Anomaly Resolution Office, and it is in the public domain.

The Department's account of what the sensor recorded is short enough to quote entire, and it is the authoritative text for this case: "The United States Indo-Pacific Command submitted a report of an unidentified anomalous phenomenon to the All-domain Anomaly Resolution Office (AARO) consisting of one minute and 59 seconds of video footage from an infrared sensor aboard a U.S. military platform in 2023. The reporter did not provide any oral or written description of the observation. Video Description: 00:00-01:59: The sensor tracks three distinct areas of contrast, maintaining their positions generally within the center of the frame. The areas of contrast appear to maintain a fixed position and orientation relative to one another. This video description is provided for informational purposes only. Readers should not interpret any part of this description as reflecting an analytical judgment, investigative conclusion, or factual determination regarding the described event's validity, nature, or significance."

This archive downloaded the released file and watched it. A monochrome infrared scene, the tracked contrasts held near the middle of the display for the full run, and heavy black redaction blocks laid over the edges of the frame where the sensor telemetry and data burn-in would normally sit. No narration, no scale reference, no readable instrument data.

The catalogue metadata is thin in a way that matters. The location field reads, in full, "(UNDISCLOSED LOCATION)", though the Department's own release manifest contradicts that and names Japan, which is dealt with below. The "Date Taken" field shows 01.01.2023, a placeholder used across these records for a file dated only to the year, not a claim about New Year's Day, and the same 0102 stub sits in the VIRIN, 230102-D-D0360-7088. Strip the paperwork away and one testable observation remains: across two minutes, three areas of contrast held a fixed position and orientation relative to one another.

What is the official explanation?

DOW-UAP-PR47 is public because of the Presidential Unsealing and Reporting System for UAP Encounters, which the Department of War abbreviates to PURSUE. It followed a directive from President Donald J. Trump, and the Department runs it with the Office of the Director of National Intelligence to "expeditiously find, review, identify, declassify and publicly release unresolved UAP-related records", in tranches posted every few weeks.

Which tranche carried this file is settled, and settled by the Department itself. Its Release 01 manifest lists the PR047 row with a Release Date of 5/8/26. That agrees with the DVIDS posting date of 05.08.2026, with the Disclosure Archive index, which groups PR46 to PR49 under "Release 5/8/26" and does not begin its "Release 5/22/26" section until DOW-UAP-PR050, and with the PURSUE page's own line that "The first tranche was released on May 8, 2026", the second following on 22 May. The CDN directory the file sits in, 2605, encodes only May 2026 and cannot separate the two releases, which is why this needed checking rather than assuming.

The word "unresolved" in the title is the Department's own, and it has defined it: "The materials archived here are unresolved cases, meaning the government is unable to make a definitive determination on the nature of the observed phenomena. This can occur for a variety of reasons, including a lack of sufficient data, and the Department of War welcomes the application of private-sector analysis, information and expertise." That is an unusual official posture. The apparatus is not asserting an identification and is not asking to be taken on trust, it is handing the file over and inviting outside work.

AARO reports to Congress under section 3373 of Title 50, United States Code. Its Fiscal Year 2025 Consolidated Annual Report, under director Dr Jon Kosloski, gives the shape of the caseload: 1,870 reports in holdings, 319 received in the year to May 2025, only 114 resolved, and 191 moved to what AARO calls its active archive, "a repository of cases lacking sufficient data to facilitate a comprehensive analysis". The report concedes that "A lack of timely and actionable sensor data continues to constrain AARO's ability to resolve cases". PR47 sits squarely inside that constraint.

Where this happened is a genuine mess, and the mess belongs to the Department rather than to its readers. The DVIDS record publishes no location at all. The machine-readable manifest that war.gov publishes alongside Release 01 does: the PR047 row carries an Incident Location of "Japan" and an alternate Video Title of "Unresolved UAP Report Japan, 2023", against a DVIDS Video ID of 1006107 that ties it unambiguously to this file. The earlier manifest already posted on the same page says the same. Two official surfaces of one record disagree, one declaring the location undisclosed and the other naming a country. The manifest also records the Incident Date as "N/A", confirming that only the year was ever published.

Handle that cell with care, because it is demonstrably ragged. In the same file the PR045 row is titled "Middle East, 2020" while its Incident Location reads "Southern United States", and its blurb credits the Department of the Air Force rather than a theatre command. The other two INDOPACOM rows are inconsistent in grain, PR046 giving "East China Sea" and PR048 only "Indo-PACOM". Metabunk member jjjdanny, reading the manifest on 12 May 2026, recorded PR047's location as "South China Sea" rather than Japan, which would mean the file was edited after publication. This archive checked both manifests as captured that same day and both read Japan, with the string "South China Sea" appearing nowhere in either, so a misreading is likelier than an edit, though neither is proved. Another member, JMartJr, offered the most useful caution: some of these cells may record where a report was sent from rather than where the incident happened. The pin on this page follows the Department's own value and sits approximately on Japan. Read it as the manifest's claim, not as a published coordinate.

The reporting command matters too. US Indo-Pacific Command is one of six geographic unified combatant commands, headquartered at Camp H. M. Smith outside Honolulu, and its area of responsibility covers, by its own description, "about half the earth's surface, stretching from the waters off the west coast of the U.S. to the western border of India, and from Antarctica to the North Pole". Reports from that theatre are rare in this tranche: of the 28 records titled "Unresolved UAP Report" in the first release, numbered PR19 to PR49, only three come from INDOPACOM, with the rest dominated by the Middle East, Syria, Iraq, the United Arab Emirates and Greece.

What did the witnesses think it was?

There is no witness account in this case, and that is not an omission by this archive. The Department states it flatly: "The reporter did not provide any oral or written description of the observation." No name, no rank, no platform type, no unit, no narrative of any kind was released with the footage, and the reporter in that sentence is an institution, the Indo-Pacific Command, not a person who looked at something and formed an impression of it.

The consequences should be stated rather than glossed. With no observer narrative there is no range, so apparent size in the frame cannot become real size, and three objects the size of airliners many miles away are indistinguishable from three the size of footballs a few hundred metres away. There is no bearing, no altitude, no heading and no speed. Nothing records whether the crew saw anything with the unaided eye, whether radar held a track, or whether the sensor was cued onto the objects by another system. Every quantity that would convert apparent motion in a gimballed sensor into real motion through the air is missing.

This is not peculiar to PR47. AARO's Fiscal Year 2025 report records that 77 of the 319 reports it received, 24.13 percent, "contained insufficient information to render an analytic conclusion on the phenomenon's morphological characteristics". Silent reporting is structural, not a quirk of one file.

The silence cuts both ways. This archive's rule is that witnesses are not put on trial, and here there is nobody to weigh at all. The absence removes the corroboration that would strengthen the case, since nobody has said the objects did anything remarkable, and it equally removes the usual sceptical lever, since nobody has claimed impossible manoeuvres or impossible acceleration. What survives is a machine recording and a paragraph written afterwards by government analysts working from that same silent file.

Is the Three Lights That Held Formation Over the Indo-Pacific (DOW-UAP-PR47, 2023) real? The two-pass assessment

Pass one, how this could be entirely ordinary. The most economical explanation for three points of infrared contrast holding a fixed position and orientation relative to one another for two straight minutes is that they are not three objects at all, but three hot spots on one airframe. Navigation lights, anti-collision strobes, engine nacelles and exhaust plumes sit in a rigid geometry by construction, and a conventional aircraft seen at distance on a coarse thermal sensor can resolve into a small cluster of separated bright patches that never change their arrangement, because physically they cannot. The Department's own wording describes an aeroplane on infrared as precisely as it describes anything exotic.

Next is a real formation of separate aircraft or drones. Militaries fly formations constantly, formation flying is defined by holding relative position, and the Indo-Pacific is one of the busiest military air environments on earth. Third is the optical-artefact family: internal reflections, ghosting and blooming inside the sensor optics produce features that move with the sensor rather than with the sky, holding a rigid geometry indefinitely. That argument, applied to earlier Navy infrared releases, belongs to named civilian analysts, principally Mick West and the Metabunk community, and this archive already carries their reading on its GIMBAL case. Fourth, hot spots on the ground or sea surface tracked from a moving platform can look like a stable arrangement of bright points that appears to move only because the aircraft is moving.

The material makes all four hard to test and none provable. On Metabunk, the staff member posting as flarkey reviewed the whole first release on 8 May 2026 and called it "more of the same", specifically "thermal videos from Reaper type UAS that show small IR features at distance with lots of parallax and limited context or a narrow FoV". Brian Dunning, in the same thread on 10 May, made the resolution point, that "The best sensors are still 640x480" on platforms of this class. What has not happened is the thing that would move this case to a disputed tier: that same community produced a concrete candidate identification for the sibling clip PR46 within hours, and for PR47 no named analyst has published an evidenced identification at all.

Pass two, what stands if the record is taken at face value. The Department reviewed this footage, cleared it through declassification, published it, and left it unresolved. That is neither an official debunk nor an endorsement. It is an admission of an unclosed file by the office whose purpose is closing them. The formation-keeping was recorded by government analysts examining the imagery, not volunteered by an excited observer, which makes it comparatively durable. If the three contrasts really were separate bodies with no airframe joining them, holding station on each other for two minutes over the Indo-Pacific, that is operationally significant whatever they turn out to be. Multiple-object formations recur in this corpus, alongside the four-object formation near Iran in PR050 and the line of contrasts over the South China Sea in PR101, but that pattern is context, not evidence about this clip.

The verdict is Unknown. There is a real, primary, officially acknowledged government record here, and the object remains officially unexplained. But the evidence is two minutes of silent, contextless, redacted infrared with no range, no scale and no witness, and the released frames do not settle whether the three contrasts were even three things. The honest position is that the strongest prosaic reading, three fixed lights on one conventional airframe, is unproven, and so is everything else.

Sources

More cases from this region: UFO sightings in Japan

Related cases

← PreviousThe Three-Pronged Object Over the Indo-Pacific (DOW-UAP-PR46, 2024) Next →The Night Airline Pilots Were Told to Manoeuvre Around Something Over Oregon (7 December 2024)