Has Something Changed in the Near-Earth Meteoroid Environment?

Since the start of 2026, the AMS has been fielding a growing number of inquiries from journalists, scientists, and the public about whether fireball activity has increased. The short answer is yes—but the details matter. We went to the data to understand exactly what has changed and, just as importantly, what hasn’t.

The AMS fireball reporting system has been in continuous operation since 2005 and reached maturity around 2016–2018, with annual Q1 event totals stabilizing in the range of 1,100–1,400 events. What follows is an analysis of Q1 data from 2011–2026, with particular attention to the 2021–2026 window where the reporting platform has been fully stable.

The Signal Is at the Top of the Distribution

The most important finding from our analysis is that the total number of fireball events is not dramatically unusual. Q1 2026’s 2,046 total events is the highest on record but only marginally above 2022 (2,037) and 2021 (1,947). If this were simply a matter of more people filing reports, we would expect a proportional increase across all witness-count thresholds. That is not what we see.

At the 25+ report threshold, 2026 has produced 61 events versus a 2021–2025 average of roughly 43—up about 42%. At 50+ reports, 2026 has 38 events versus an average of 18—more than double. And at 100+, the count of 14 is twice the average of 7. The signal gets stronger as the threshold rises, which is the hallmark of a genuine physical change in the incoming material, not a reporting artifact.

March 2026: The Distribution Shifted Upward

Breaking Q1 2026 down by month reveals that January was mildly elevated, February showed a noticeable increase at the upper end, and March is where the signal becomes unmistakable.

The key observation is that the base layer of activity (10–24 reports) is essentially unchanged at 21 events, virtually identical to every other year. What has changed is that a large fraction of events that would normally draw 25–49 witnesses instead drew 50, 100, or even 200+ witnesses. The distribution didn’t broaden—it shifted upward. Almost half of all March 2026 events with 10+ reports were seen by 50 or more people.

The March average witness count per event was 142.7—nearly three times the next-highest March on record (49.4 in 2021). Even removing the massive 3,229-report German event on March 8, the remaining 41 events average roughly 67 reports each, still more than double the historical norm.

Physical Characteristics: Sonic Booms Confirm Larger Objects

If the fireballs were simply being seen by more people due to favorable conditions, we would not expect changes in the physical characteristics reported by witnesses. But the data shows an elevated rate of delayed sound reports—sonic booms reaching the ground—which requires objects that penetrate deep enough into the atmosphere to produce pressure waves.

In Q1 2026, 30 of 38 events with 50+ reports (79%) produced sonic booms, placing it in the top tier historically. But what makes 2026 unique is the combination: prior high-sound years like 2021 and 2023 had elevated percentages but moderate event counts. In 2026, both the rate and the absolute count are high. Thirty large fireball events producing audible booms in a single quarter means roughly one every three days.

Meanwhile, the total number of individual long-duration sighting reports (witnesses reporting fireballs lasting 4+ seconds) reached 1,693 in Q1 2026—more than 2.5 times the previous high of 651 set in 2021.

The Remarkable Events of March 2026

While the aggregate statistics tell one story, the individual events of March 2026 have been extraordinary. Beginning with a massive European event on March 8 and accelerating through a sustained wave from March 11–24, the AMS logged an unprecedented concentration of major fireballs—many producing confirmed meteorite falls.

Radiant Analysis: Enhanced Activity From Known Sources

When fireballs originate from a common parent body or debris stream, their radiants—the apparent points in the sky from which they travel—cluster together. We computed radiants for all 61 trajectory-resolved events in Q1 2026 and compared them to the 2021–2025 baseline. The results reveal meaningful clustering in two regions of the sky.

Anthelion concentration. The Anthelion sporadic source—the region of the sky opposite the Sun, centered near RA 150–200°, Dec -5° to +20° during Q1—is the dominant source of sporadic fireballs in every year. But in 2026, activity from this region has approximately doubled. Twelve events fall in this tight zone in 2026, compared to 1–6 in prior years. The density of 9.6 events per 1,000 square degrees is twice the previous maximum. This cluster includes several of the quarter’s largest events, including the March 11 France fireball (236 reports), the March 9 East Coast event (282 reports), and the March 10 France event (145 reports).

High-declination enhancement. Eleven events in 2026 originated from radiants above Dec +70°—more than double the next-highest year (2021 with 5). This cluster includes the Ohio eucrite (Dec +77°), the California fireball (Dec +81°), and the Feb 11 Ohio event (164 reports, Dec +80°). High-declination radiants correspond to objects on steeply inclined orbits relative to the ecliptic plane. An enhancement in this population is unusual and warrants further study.

The two eucrite falls remain unrelated. Despite both being rare achondritic eucrites, the Germany (RA 33°, Dec -11°) and Ohio (RA 112°, Dec +77°) events have an angular separation of 98.2 degrees—they came from opposite parts of the sky on unrelated orbits. Two eucrite falls in nine days is statistically remarkable, but they are not from a common stream.

Important caveat. These radiants are computed from witness-based trajectory solutions, not instrumental camera data. The typical uncertainty in witness-derived radiants is 10–20 degrees, meaning the apparent clustering could be tighter or looser than shown. Instrumental confirmation from allsky camera networks would substantially improve the precision of these results. The raw radiant data for all 255 events used in this analysis is available for download below.

What We Can Rule Out

Increased reporting or smartphone adoption. The AMS reporting platform has been mature since 2016–2018. The total event count for Q1 2026 is only marginally above recent years. The anomaly exists only at high witness-count thresholds—the opposite of what a reporting effect would produce.

A new meteor shower. No major showers are active in Q1 (the next is the Lyrids in April). The enhanced activity concentrates around the Anthelion sporadic source and high-declination radiants rather than a novel radiant position. This is consistent with an amplification of known sporadic populations rather than a new stream.

The “February fireballs” seasonal effect. The well-documented February fireball phenomenon—thought to be related to the Anthelion radiant reaching peak evening elevation—typically produces a 10–30% increase in bright sporadic meteors. Q1 2026’s 50+ event count is more than double the baseline, and the signal is strongest in March, not February. The Anthelion enhancement we observe may be related to this phenomenon but substantially exceeds its historical magnitude.

Time-of-day or geographic bias. The evening/overnight/morning distribution of events is statistically normal. The surge is global, with major events in both North America and Europe.

What We Cannot Yet Rule Out

AI-driven reporting amplification. One factor that has changed since 2023 is the widespread adoption of AI assistants. When someone witnesses a fireball today, they may ask ChatGPT, Siri, or Google’s AI “I just saw a fireball—where do I report it?” and be directed to the AMS. This would inflate witness counts per event without changing the actual number of fireballs—which is, notably, the exact pattern we observe: normal total event counts but elevated reports per event at the high end. We cannot quantify this effect with the data currently available, but it is a plausible partial explanation for the upward shift in the witness-count distribution. It would not, however, account for the elevated sonic boom rates or the recovered meteorite falls.

The Uncomfortable Questions

Understandably, dramatic fireballs over populated areas prompt speculation. We address two common questions directly.

Are these alien in origin?

No. Every fireball in the AMS database with sufficient trajectory data is consistent with objects on heliocentric orbits—material orbiting the Sun that intersects Earth’s path. Entry velocities, entry angles, and orbital characteristics match the known sporadic meteoroid complex. The recovered specimens from Ohio and Germany are achondritic eucrites with mineral compositions formed over billions of years on differentiated asteroids. These are rocks from the inner solar system. There is no evidence of anomalous trajectory behavior, controlled flight, or non-natural composition.

Could any of this activity be artificial?

For the events with recovered specimens—specifically the Ohio and Germany eucrites—the mineralogical evidence is definitive. These are natural extraterrestrial rocks with formation histories traceable to specific parent body types. However, not every fireball in the dataset has recovered meteorites, and the AMS trajectory computation system relies on witness reports, not instrumental tracking. For the events we cannot fully characterize, we simply lack sufficient data to make definitive statements about every single one. This is precisely why expanded instrumental coverage matters.

What We Still Need to Learn

The most honest answer to “why is this happening?” is that we don’t fully know. The data points to a genuine enhancement in the sporadic fireball background at the large-object end of the size distribution. Whether this represents normal statistical variance at the tail of the distribution, an uncharacterized debris population, or something else entirely will require continued monitoring and further analysis. Here is what would help:

Expanded allsky camera coverage. At the time of the Ohio fireball, there was only one AMS-affiliated camera station in the entire state of Ohio—and it was offline. A national allsky camera network would provide instrumental trajectory, velocity, and orbit data for every bright fireball, independent of witness reports. It would allow us to definitively classify each event and determine its origin. The case for such a network has never been stronger: in a quarter where the fireball rate at the 50+ witness level doubled, we need instruments that can tell us exactly what is entering our atmosphere and where it came from.

Laboratory analysis of recovered specimens. The Ohio eucrites and any confirmed Houston specimens should be studied for cosmic ray exposure ages and orbital history. If the two eucrite falls share a common cosmic ray exposure age, that would indicate a recent disruption event on the parent body—even though their radiants differ.

Statistical modeling. A rigorous Poisson analysis of the full 2005–2026 dataset would quantify whether the Q1 2026 surge falls within expected variance or is anomalous at high confidence.

Cross-network sensor correlation. Data from NASA’s All-Sky Fireball Network, NOAA’s GOES satellite lightning mapper, infrasound arrays, and Doppler weather radar should be systematically cross-referenced with the AMS witness database. Several Q1 2026 events were detected by multiple sensor types; this multi-modal approach is the gold standard for characterizing bright fireball events.

What We’re Saying—And Not Saying

This is not evidence of an impact threat. The objects involved range from pebble-sized to a few meters across and are part of the normal continuum of material that Earth encounters. None posed a danger beyond localized effects (sonic booms, the rare roof strike in Houston).

What this is, is a measurable change in the AMS fireball data that we do not yet fully understand. After years of stable baseline activity, something appears to have shifted in Q1 2026, and the signal is consistent across multiple metrics: witness counts, sonic boom rates, long-duration sighting volume, and the distribution of event sizes. Whether this reflects a genuine change in the near-Earth meteoroid environment, an amplification of reporting through AI and social media, or some combination of both—we cannot yet say definitively. What we can say is that the question deserves both public awareness and scientific attention.

Supporting Data

In the interest of transparency and reproducibility, the radiant dataset used in this analysis is available for download. It includes event ID, date, year, witness count, computed RA and declination (degrees), azimuth, entry angle, ground-track distance, and start/end coordinates and altitudes for all 255 events from Q1 2021–2026 with 25+ witness reports and valid trajectory solutions.

Radiants were computed using the same methodology as the AMS event processing pipeline: haversine ground distance between trajectory start and end points, entry angle from altitude difference, compass bearing from end point to start point, and conversion to equatorial coordinates (RA/Dec) via PyEphem’s observer model. The start altitude is a fixed assumption of 80 km in the trajectory solver; the end altitude is computed. Events where start and end coordinates were identical (degenerate solutions) were excluded. Note that these radiants are derived from witness-based trajectory fits and carry typical uncertainties of 10–20 degrees. We welcome independent analysis.