On a clear night in late 2024, U.S. government satellites recorded a fireball over the western Pacific that released energy equivalent to roughly 400 tons of TNT. It was not unusual by itself. What was unusual was how many events like it had been piling up. By early 2026, researchers analyzing the federal fireball archive noticed that recent reporting rates had climbed to 3.9 standard deviations above the historical mean, a statistical outlier so extreme it should occur by chance less than one time in 10,000. As of June 2026, neither NASA, the U.S. Space Force, nor any peer-reviewed study has offered a definitive explanation.
The data behind the number
The dataset driving the conversation is the Fireball and Bolide Reports archive hosted on the federal open-data portal catalog.data.gov. Maintained through NASA’s Center for Near-Earth Object Studies at the Jet Propulsion Laboratory, it logs each detected event by date, time, geographic coordinates, altitude, velocity components, optical radiated energy, and a calculated total impact energy derived from an empirical conversion formula. Records are available in JSON, CSV, XML, and RDF formats, and the JPL API documentation provides technical specs for programmatic queries.
The sensors feeding this archive are U.S. government satellites originally built for missile warning and nuclear detonation detection. For decades, the data they collected on natural atmospheric impacts stayed classified. That changed when the Space Force released decades of bolide observations to NASA for planetary defense research, giving civilian scientists access to instrument-grade measurements that had previously been locked behind national security restrictions.
A second, independent detection channel now supplements the military sensors. The Geostationary Lightning Mapper aboard NOAA’s GOES-16 and GOES-17 weather satellites, operational since 2017 and 2018 respectively, was designed to track lightning in near-real time. But researchers at NASA’s Ames Research Center recognized that GLM’s wide-field optical sensors were also sensitive enough to capture the brief, intense flashes of incoming bolides. They built a machine-learning pipeline to extract bolide lightcurves from GLM data and began publishing detections to a NASA-hosted portal. The method was described in a preprint on arXiv and a peer-reviewed version published in Icarus.
Because GLM and the defense satellites operate on entirely different platforms with different design goals, they offer a partial cross-check. When both systems record the same event, researchers can compare energy estimates and timing. When only one system flags an event, the gap can reveal differences in sensitivity, coverage area, or detection thresholds.
Why the spike might not be what it looks like
The core uncertainty is deceptively simple: are more rocks actually hitting the atmosphere, or are better instruments catching events that older hardware would have missed?
A peer-reviewed analysis published in Monthly Notices of the Royal Astronomical Society examined both CNEOS sensor data and GLM bolide detections and flagged potential inconsistencies in the record tied to changes in satellite coverage, sensor sensitivity, and data-release policies over time. The finding raises a concrete possibility. If the addition of GLM’s expanded optical detection capability after 2017 inflated apparent event counts in the combined record, the spike could be an artifact of better observation rather than a real increase in incoming debris.
The defense satellites themselves present a similar problem. If newer generations of these spacecraft can detect fainter flashes or lower-energy impacts than their predecessors, then the historical baseline of “normal” activity may be artificially low. A 3.9-sigma deviation measured against a suppressed baseline would overstate the true anomaly. Without a transparent, instrument-by-instrument calibration history, which remains unavailable due to classification, researchers can only approximate how detection thresholds have shifted over the decades.
Eyewitness reporting adds another wrinkle. A study archived in PubMed Central evaluated how meteor trajectories derived from witness accounts submitted to the American Meteor Society compare with instrument-based measurements. The researchers found that trajectory accuracy depends heavily on the number of independent reports for a given event, meaning a surge in public sightings does not automatically signal a surge in actual fireballs. More people carrying smartphones, more awareness of reporting platforms, and more media coverage of any single dramatic event can all inflate filing counts without any change in the sky overhead.
What natural explanations would look like
If the increase is real and not an observational artifact, planetary scientists would look for a physical mechanism. Earth periodically passes through debris streams left by comets, producing predictable meteor showers like the Perseids and Geminids. But those showers involve small particles, typically sand-grain to pebble-sized, not the meter-scale objects that produce the bright fireballs tracked by CNEOS.
A genuine uptick in large fireballs could theoretically result from the breakup of a previously unknown asteroid that scattered fragments into Earth-crossing orbits, or from a long-period comet whose debris trail Earth has only recently begun to intersect. Survey telescopes like the Catalina Sky Survey and the ATLAS system, which scan the sky nightly for near-Earth objects, have not publicly reported a corresponding increase in newly discovered objects on collision-relevant trajectories. That silence does not rule out the possibility, since small, dark fragments a few meters across are extremely difficult to spot before they arrive, but it does mean there is no corroborating signal from the discovery side of planetary defense.
Another possibility involves orbital mechanics and timing. The population of near-Earth objects is not static; gravitational interactions with Jupiter and the inner planets continuously shuffle orbits. Small shifts in the orbital distribution of meter-class objects could, over years, funnel more of them into Earth-crossing paths. But modeling that process at the resolution needed to explain a short-term spike would require detailed orbital data for a population that is, by definition, mostly untracked.
What would settle the question
Researchers who have examined the data publicly point to several steps that could resolve the ambiguity. The most direct would be for the Space Force or the National Reconnaissance Office to release a calibration history for the satellite sensors that feed CNEOS, documenting how detection thresholds changed with each hardware generation. That would allow analysts to normalize the historical record and determine whether the baseline was artificially low in earlier decades.
A second approach would isolate the CNEOS dataset into pre-GLM and post-GLM subsets and cross-reference each with independent ground-based networks, such as the European Fireball Network or the Desert Fireball Network in Australia, that have operated with relatively stable instrumentation over long periods. If the spike appears only in the satellite data and not in ground-based records, the case for an observational artifact strengthens considerably.
A third test is available to anyone with basic data skills: download the CNEOS archive from data.gov, filter for bolides above an energy threshold high enough that even older satellites would have detected them, and chart annual counts. If the spike persists at high energies, it is harder to attribute to improved sensitivity. If it disappears and the deviation is driven entirely by smaller, fainter events appearing in recent years, the instrument-upgrade explanation gains ground.
Until one of these analyses is completed and published, or until NASA or the Space Force issues an official assessment, the 3.9-sigma figure sits in an uncomfortable middle ground. The data genuinely show an unusual clustering of bright fireballs. But the observational system that produced the data has itself been evolving, and in planetary defense, where rare but high-consequence events drive policy, mistaking better detection for a worsening threat could be just as costly as missing a real change in the sky.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
