AARO Information Paper: Starlink Satellite Flares Mistaken for UAP

In March 2024, staff from the Pentagon’s UAP office travelled to the High Plains of western Nebraska, set up a consumer camera and waited for the sky to do something strange. It did, when and roughly where their calculations said it would. This 18-page information paper explains how they knew.

What this record is

This is “Correlations of Starlink Satellite Flaring with UAP Observations,” an Information Paper from the All-domain Anomaly Resolution Office (AARO) dated December 2024. The National Archives catalog lists it as the “AARO Satellite Flaring Paper.” It has 18 pages and 16 figures, with worked tables, equations and a reference list that begins with work by the astrophysicist Jonathan McDowell on the Starlink population and a 1959 Naval Research Laboratory paper on seeing the Vanguard satellite. The Department of Defense prepublication office cleared it for open publication on 23 January 2025, and it carries AARO case number 330UAP000027, released in full. A footnote says the focus on Starlink implies no endorsement and that the principles apply to any similar constellation.

What it says

The paper begins with scale. It says there are nearly 10,000 artificial satellites in low Earth orbit, a number expected to grow several times over, and more than 6,700 of them were Starlink satellites at the end of November 2024. It then distinguishes two kinds of reflection. Diffuse reflection from rough surfaces spreads sunlight widely, so a satellite looks like a moving star for up to several minutes. Specular reflection, or glint, from mirror-like surfaces sends light into a narrow cone, and an observer inside that cone sees a flash “orders of magnitude brighter than starlight” that lasts only a moment. On a Starlink satellite, the paper says, the underside of the bus, with its mirrored panels and flat antennas, is a very efficient specular reflector.

It separates flares from “Starlink trains.” Trains are the lines of lights seen for several days after a launch, while dozens of satellites climb in a streamlined attitude with their solar panels flat. Flares come later, once the satellites reach operating altitude, nominally 550 km, and turn their buses to face the ground. The paper warns that several satellites flaring at once can look like “spinning lights, small glowing orbs that disappear and reappear, or tracing out geometric shapes such as triangles.” It notes that flaring itself is not new and was well documented with the Iridium satellites in the late 1990s.

The method rests on geometry. For a 550 km orbit and reflections up to 2 degrees off axis, Starlink flares are generally visible when the Sun is roughly 38 to 46 degrees below the horizon, and they appear within about 10 degrees of the horizon. The paper walks the reader through using free Sun calculators to find the times and compass directions of that “flare window.” It then tests the method. AARO staff travelled to Sidney, Nebraska, from 8 to 12 March 2024, choosing it for its altitude of about 4,000 feet and its flat terrain. They predicted a window after sunset on 10 March between 10:22 and 11:20 p.m., and before sunrise on 11 March between 2:44 and 3:43 a.m. They then photographed flares at 10:38 p.m. (azimuth 309.6 degrees, elevation 3.2 degrees) and at 3:13 a.m. (azimuth 47.4 degrees, elevation 3.7 degrees), both inside the predicted ranges. The camera was a Canon EOS Rebel T7 set to 10-second exposures, which turned the flares into short streaks.

For people in aircraft, altitude pushes back the horizon, and the paper shows how to calculate the equivalent ground position. It applies this to a real report. On 9 October 2022, at about 3:50 a.m., an airline pilot flying east at 35,000 feet near Gallup, New Mexico, reported to the Federal Aviation Administration “multiple lights moving in different directions, left of the constellation Leo.” Adjusted for altitude, the pilot’s effective viewpoint was near Cimarron, New Mexico, where the Sun was 38.5 degrees below the horizon and the predicted flare window lay about 9 degrees up at an azimuth of 65.6 degrees. AARO estimated the pilot’s sighting at an azimuth of about 63.3 degrees and an elevation of about 5.7 degrees. A planetarium reconstruction showed Starlink satellites crossing that area in different directions, and AARO concluded the lights were “very likely Starlink and other satellite flares.”

Context

The paper connects to several pages already on this site. The Gallup report appears as the October 2022 Gallup record and, under the pilot’s effective viewpoint, as the Cimarron record. The Nebraska photographs appear as the Sidney, Nebraska record. Those entries were extracted from single pages of this paper when the Department of War released it in May 2026. Read in context, the Sidney pages show a planned photography trip by AARO staff, not a sighting. Civilian reporting services already warn witnesses about Starlink. The National UFO Reporting Center, for example, shows Starlink examples on its reporting page, as AARO’s own sky-view study notes. AARO’s paper goes further than a warning. It gives observers a method they can check for themselves with free tools.

Status

This paper explains a known phenomenon and applies it to one case. The explanation is well founded. The optics are standard, the predictions were confirmed in the field, and the Gallup analysis is specific enough to check. The Gallup result is still an inference. The pilot gave no azimuth or elevation, so AARO worked them out from the phrase “left of the constellation Leo.” The paper’s broader claim is limited. It says satellite flares “can explain some” reports, not all of them.

Sources

• National Archives catalog entry: https://catalog.archives.gov/id/499915959
• PDF of the record: https://catalog.archives.gov/medialz/electronic-records/rg-615/493468580/027_Satellite_Flaring_Paper.pdf
• Series: Records from the Office of the Secretary of Defense Relating to Unidentified Anomalous Phenomena: https://catalog.archives.gov/id/493468580
• National Archives, Record Group 615 UAP records: https://www.archives.gov/research/topics/uaps/rg-615