“On the day after a nuclear test, you saw transients on almost 19 percent of those days,” says Stephen Bruehl of Vanderbilt University. That number compared to just 11 percent on non-test days was one of the most surprising results of a recent examination of archival astronomical data that uncovered an unexpected statistical connection between mid‑20th‑century nuclear detonations and momentary flashes of light recorded across the skies.
This study focuses on the northern sky explored through the Palomar Observatory Sky Survey (POSS‑I) between 1949 and 1957, using the 1.2‑meter Schmidt telescope in California. Over a period of approximately 2,000 glass photographic plates, each exposed for around 50 minutes, astronomers developed a record of the northern sky in unprecedented detail. These plates were treated with light‑sensitive emulsions, making them the standard medium for imaging before the development of digital detectors. They were digitized decades later, enabling researchers to re-examine them using up-to-date image-processing algorithms. Alongside stars and galaxies, the team identified upwards of more than 107,000 “transients”-star‑like points that appear in one plate but vanish in earlier and later exposures.
Historically, such anomalies were often dismissed as defects: dust specks, emulsion flaws, or cosmic ray hits. Yet the Vanishing & Appearing Sources during a Century of Observations (VASCO) project, led by Beatriz Villarroel of Stockholm University’s Nordita, has shown that many transients have point spread functions consistent with real celestial sources. Some even appear in multiples on a single plate, aligned in ways that natural astrophysical events rarely produce.
To explore possible causes, the researchers compiled a dataset covering 2,718 days of POSS‑I observations. They compared transient dates with all known above‑ground nuclear test dates for the United States, Soviet Union, and Great Britain, and also compared them with witness reports of unidentified anomalous phenomena (UAPs) from the comprehensive UFOCAT database. Statistical tests revealed that the transients were 45 percent more likely to occur within a “nuclear test window”-defined as the day before, the day of, or the day after a detonation. The most pronounced effect came one day after a test, with a 68 percent higher likelihood of transient occurrence.
The team also analyzed UAP reports. Though the sightings were frequent-on almost 90 percent of study days-the number of transients increased with the number of independent UAP reports. For each additional UAP sighting on a given day, transient counts increased by 8.5 percent. Where nuclear test windows and UAP reports coincided, their effects were additive: days with both showed more than double the transient activity compared to days with neither.
From an engineering and physics perspective, there are indeed a number of mechanisms that could plausibly link nuclear detonations to transient phenomena. For one thing, above-ground tests release copious amounts of gamma radiation and charged particles into the atmosphere, which could in turn produce Cherenkov radiation – visible light emitted when high-energy particles travel through air faster than its phase velocity of light. Nuclear blasts also create electromagnetic pulses, and metallic debris and radioactive dust are injected into the upper atmosphere. High-altitude particles or flat reflective fragments could briefly glint in sunlight, thereby mimicking starlike points in long-exposure astronomical images. That many transients were absent in Earth’s umbral shadow – where solar reflections cannot occur-supports the notion that at least a fraction of them are indeed caused by sunlight reflecting off orbiting objects.
Some dates offer tantalizing case studies. Multiple bright transients were recorded by POSS‑I plates on July 19 and July 27, 1952; these nights coincided with the famous Washington, D.C., UAP wave, when radar and visual sightings persisted for hours. If such objects were in geostationary orbit about 35,000 kilometres above Earth they would appear stationary during a 50‑minute exposure, producing a point source rather than a streak. Lower‑altitude objects, meteors, or atmospheric effects would typically leave trails.
The authors eliminated several possibilities. Fallout contamination of plates produces diffuse fogging, not sharp point spread functions. Observer bias is implausible; the existence of transients was unknown at that time, and nuclear test schedules were not public. The timing, peaking the day after the tests, speaks against immediate effects due to debris and is most simply interpreted in terms of either lingering atmospheric phenomena or objects arriving in response to the detonations.
While the correlations are small in magnitude, they are statistically robust due to the large data set. Noise from automated transient detection and variability in UAP report reliability likely dilute the signal. Future efforts can utilize AI to improve transient identification, inspect other photographic plate archives for similar patterns, and apply the methodology on modern geostationary satellites to calibrate possible reflection signatures.
Whether these mid‑century flashes were atmospheric artifacts, human‑made objects in high orbit, or something more exotic, the study shows that careful statistical analysis of historical astronomical data can uncover subtle links between terrestrial events and celestial observations links that may help unravel decades‑old mysteries.
