In September 2026, the VASCO project team, led by astrophysicist Beatriz Villarroel, presented several important results at once that significantly advance the investigation of unusual short-lived flashes on old photographic plates. These events, recorded even before the launch of the first artificial satellite, have been of interest to researchers for several years, and new works add significant details to them.
Dear everyone, Some new and very exciting news from the VASCO team! First, our machine-learning paper has been accepted for publication following peer review and is now in press. In this paper, we used machine learning to address some of the criticism concerning plate defects and
First, it became known that a paper applying machine learning methods has passed peer review and been accepted for publication (it is currently in press). In it, scientists cleaned the sample of potential emulsion defects. As the data became cleaner, key statistical correlations did not weaken, but rather strengthened. This applies to both the deficit of flashes in the Earth's shadow and the connection to nuclear test dates.
The second result relates to determining the altitude of the objects that could have produced these flashes. On September 4, 2026, a preprint titled 'Modelling Palomar Transients: Constraints from Reflection Geometry and Orbital Altitude' (arXiv:2609.05105) appeared on arXiv. Using Earth shadow geometry and modeling, researchers estimated that the sources are located at altitudes ranging from approximately 20 000 to 35 786 kilometers above the Earth's surface. The upper limit corresponds to geostationary orbit. The objects are noticeably clustered closer to the equator.
Modeling also allowed for the estimation of physical properties. Judging by the nature of the sunlight reflection, these are small flat surfaces with high reflectivity—ranging from a few centimeters to about three meters. The flashes last fractions of a second (a characteristic time of about 320 milliseconds), and the objects themselves may be slowly rotating or abruptly changing orientation. The authors are considering various working models, both natural and otherwise, that could fit the observations.
The third direction is geographical. On September 8, 2026, a second preprint was released—'Earth-Projected Clustering of Historical Optical Transients in the Palomar Observatory Sky Survey-I (POSS-I)' (arXiv:2609.09461). Projections of the pre-Sputnik flash positions revealed peculiar 'hotspots': the Sedona area (Arizona), White Sands/El Paso, Tampico, and several other locations, including Pacific waters west of southern Mexico and the southern Gulf of Mexico. A connection to nuclear testing is also evident in this data: flashes during Pacific tests were more frequently projected onto the ocean, while those during Nevada tests were primarily projected onto the southwestern United States.
Researcher Dennis Åsberg, commenting on these results, noted that as the data is cleaned, the picture does not fall apart but becomes more consistent: altitude, equatorial concentration, specular reflections, geographical clusters, and the connection to nuclear testing are beginning to form a unified outline. Exactly what produced these flashes before the appearance of the first satellites remains an open question. However, it is becoming increasingly difficult to ignore the accumulated patterns.
The work continues. The team emphasizes that the conclusions are still cautious and require further verification, including modern observations. Nevertheless, the new preprints and the paper accepted for publication show that old photographic plates continue to reveal details that until recently seemed almost imperceptible. And the more carefully researchers clean the data, the more clearly what once flickered in the night sky long before the start of the space age begins to emerge.
