Paper Flags April 1 South Atlantic Fireball as Leading Post‑2018 Interstellar Meteor Candidate

A new arXiv preprint argues that an April 1 fireball over the South Atlantic may be the strongest post-2018 interstellar meteor candidate in NASA JPL’s public CNEOS fireball database. But the claim comes with a central caveat: It relies on U.S. government sensor data that CNEOS, the Center for Near-Earth Object Studies, says it does not independently verify or reanalyze.

If confirmed, an interstellar meteor would be a rare piece of matter from outside the solar system. But this case is not an established detection. It is a preprint — a paper posted online before peer review — and its conclusion depends heavily on whether the cataloged speed and direction for the event are accurate enough to support an extrasolar origin.

The manuscript, “A High-Likelihood Polar Interstellar Meteor Candidate,” was posted to arXiv as arXiv:2606.04379v1 and submitted June 3 by Richard Cloete and Abraham “Avi” Loeb, the Harvard astrophysicist. In the abstract, the authors write: “We report a newly identified polar interstellar meteor candidate, labeled polarIM, detected on 2026-04-01 02:13:14 UTC at latitude -41.9°, longitude -54.7°, and altitude 90.5 km over the South Atlantic Ocean, east of Argentina.”

According to the paper, the event’s reported Earth-fixed velocity vector was (+3.6, -34.6, +59.8) kilometers per second. Cloete and Loeb said they converted that measurement into a heliocentric trajectory — a path around the sun — by transforming it to an inertial geocentric state, removing Earth’s gravitational acceleration with a two-body hyperbolic model, and adding Earth’s heliocentric velocity from JPL Horizons. Their reported result was a heliocentric speed of 51.73 kilometers per second, a heliocentric excess speed of 30.00 kilometers per second, positive heliocentric specific energy of +450.1 square kilometers per square second, and an inclination of 89.4 degrees. The paper also says the meteor’s final polar, or z, velocity component of +47.09 kilometers per second exceeded the local solar escape speed of 42.14 kilometers per second.

The authors then tested how sensitive that conclusion was to measurement uncertainty. Using 1,000,000 Monte Carlo realizations and an empirical post-2018 CNEOS error model from Peña-Asensio et al. (2025), with stated uncertainties of 0.55 kilometers per second in speed, 1.35 degrees in right ascension and 0.84 degrees in declination, they reported that none of the simulated cases produced a bound orbit around the sun. As the abstract puts it, “No realization yields a bound heliocentric orbit, giving a statistical confidence on the interstellar fraction of >99.9997%.” They also describe polarIM as the highest-margin post-2018 candidate in the CNEOS catalog under their analysis.

That conclusion, however, rests on data that CNEOS itself warns should be treated carefully. On its Fireballs page, CNEOS says the database is derived from U.S. government sensor detections and that “CNEOS does not independently verify or reanalyze these events.” It adds that the data “should be used with appropriate caution.” That warning is the key issue here: The paper’s interstellar claim depends on the underlying measurement and on the adopted error model being reliable enough to rule out a solar system origin. Earlier claims that some CNEOS fireballs were interstellar have been contested in the literature on those grounds.

Loeb also posted a public summary on Medium repeating the event details and conclusions. In both the paper and the public note, the authors call for independent validation, including improved trajectory reconstruction, additional uncertainty testing, more detailed back-integration of the orbit and a search for other observations from that date. Practical follow-up appears limited. The paper says the event’s energy was small, about 0.086 kilotons of TNT equivalent, and that it occurred high in the atmosphere, making any fragment recovery difficult.