Pith. sign in

REVIEW 5 minor 18 references

Early 2026 AMS fireball reports match the long-term growth trend with no statistical surge or other anomalies.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · grok-4.5

2026-07-11 09:14 UTC pith:6AGVERC7

load-bearing objection Clean, well-executed nulls on the AMS 2026 surge claims plus a usable Poisson-regression primer; the central count results hold up.

arxiv 2607.05071 v1 pith:6AGVERC7 submitted 2026-07-06 astro-ph.EP astro-ph.IM

A cornucopia of null results: A statistical analysis of fireballs reported to the American Meteor Society

Althea V. Moorhead This is my paper
classification astro-ph.EP astro-ph.IM
keywords Poisson regressionfireballsAmerican Meteor Societymeteor ratesradiant distributionnull resultsgeneralized linear modelsquantile residuals
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The American Meteor Society announced an apparent surge of large fireballs in the first quarter of 2026, citing higher counts, radiant clustering, and delayed-sound reports. This paper treats the public count data as a case study for Poisson regression and related residual and multiple-comparison tools. It finds the observed numbers sit squarely on the multi-year linear rise in reports, ordinary monthly variation, and the expected power-law fall-off with reporting threshold. Residual checks and Bonferroni-adjusted tests reveal no outliers in any quarter or month, no shift in size distribution, no excess of delayed-sound events, and no change in radiant distribution relative to prior years. The same analysis finds little support for the long-standing lore of elevated “February fireballs.” The result is both a null finding on the claimed surge and a practical demonstration that count-based statistical methods can settle whether reported meteor activity is unusual.

Core claim

Poisson regression of AMS fireball counts shows that first-quarter 2026 activity is consistent with a steady linear increase over years, ordinary seasonal variation, and a power-law dependence on the number of reports per event; quantile residuals and Bonferroni-corrected tests detect no anomalous quarters, months, size distributions, delayed-sound fractions, or radiant distributions, and February rates are unremarkable relative to the monthly average.

What carries the argument

Poisson regression (generalized linear model for independent event counts) with year or month as predictors, log reporting threshold, log-link mean functions, quantile residuals, and Bonferroni-adjusted significance thresholds for outlier and multiple-comparison control.

Load-bearing premise

Fireball event counts are independent and Poisson-distributed around a mean completely captured by year, season, and reporting threshold, so unmodeled weather, media, or misidentification effects do not hide real anomalies.

What would settle it

A future quarter or month whose quantile residual, after refitting the same year-plus-threshold model, exceeds the Bonferroni-adjusted normal threshold, or a statistically significant interaction term showing a new year has a different size-distribution slope.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Resource planners can forecast expected AMS report volumes from a simple closed-form expression in year and report threshold.
  • Claims of unusual meteor activity must be tested against multi-year trends rather than short recent averages.
  • February is not elevated in this dataset; November shows the highest average rates.
  • Poisson models with residual-deviance checks are practical for sparse meteor-count data that previously lacked formal rate analysis.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same residual-and-Bonferroni pipeline could be applied to other citizen-science fireball archives to isolate weather or media biases from true rate changes.
  • Continued growth in reports will make future anomalies easier to detect, yet separating natural rate shifts from observational artifacts will remain the harder problem.
  • The recovered power-law index on report numbers may ultimately constrain the brightness distribution of fireballs once selection effects are modeled.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

0 major / 5 minor

Summary. The paper re-examines AMS claims of a 2026 Q1 fireball surge (and related claims about size distribution, March excess, delayed sound, and radiant clustering) using Poisson GLMs on website-scraped event counts (2011–2026). After a short primer on GLMs, quantile residuals, Bonferroni corrections, and SCE coordinates, the authors fit models with year (or fractional year), report-threshold bins, and quarter/month intercepts (Eqs. 9, 11, 16). Residual deviance, quantile residuals (Figs. 6, 8), and Bonferroni-adjusted outlier thresholds show no anomalous quarter or month; an explicit 2026 interaction (Eq. 14 / Table 5) is non-significant; Fisher exact tests find no change in delayed-sound fraction; and 2-D KS tests find no change in radiant distribution. Two claims are left untested for lack of usable data. The paper also finds no support for elevated February rates in the AMS reports.

Significance. If the nulls hold, the paper supplies a clear, reproducible counterweight to a public AMS claim and a practical template for count-based meteor statistics. Strengths include explicit model assumptions, residual and dispersion checks, multiple-comparison control, an independent data scrape (with API–website caveats documented), and a supplementary Markdown analysis file. The primer on Poisson regression and SCE coordinates is useful for a field that underuses these tools. The result is incremental rather than transformative, but it is falsifiable, well-scoped, and immediately relevant to resource planning for fireball analysis.

minor comments (5)
  1. In §3.2.2 the Bonferroni residual cutoff is given as z_test = 3.71 while Fig. 6 caption states ±3.99; reconcile the numerical value and the α_test formula used.
  2. Table 5 labels the 2026 interaction coefficients β_2026 and β_rep,2026, whereas Eq. (14) uses Δ notation; align the symbols for readability.
  3. Fig. 9 would benefit from a brief note on sample size per year and on whether the plotted radiants are apparent or geocentric, given the uncertainty discussion in §3.5.
  4. A short sentence in the conclusions quantifying residual overdispersion (deviance vs residual d.f. in Tables 3–6) would help readers judge how much unmodeled variance remains after the chosen predictors.
  5. Typos / wording: “we readers to §7.5” (§2.3); “are are hidden” (§2.1); ensure consistent hyphenation of “first-quarter” / “Q1”.

Circularity Check

0 steps flagged

No circularity: standard Poisson GLM hypothesis tests on public AMS counts; fitted coefficients construct prediction intervals and residuals, not the claims under test.

full rationale

The paper re-analyzes publicly posted AMS fireball counts and radiants with ordinary Poisson GLMs (Eqs. 7–9, 11, 14, 16), residual deviance checks, quantile residuals (Figs. 6, 8), Bonferroni-adjusted outlier thresholds, Fisher exact tests, and a 2-D KS test. Coefficients (β_yr, β_rep, monthly intercepts, 2026 interaction terms) are estimated from the data and used only to form expected rates, prediction bands, and p-values under the null that 2026 follows the same trend; none of the five tested AMS claims is recovered by construction from those fits. The single self-citation (Moorhead et al. 2026) is an incidental example of count regularization and is not load-bearing. No equation reduces to its own input, no uniqueness theorem is imported, and no ansatz is smuggled via citation. The analysis is therefore self-contained against the external AMS claims it evaluates.

Axiom & Free-Parameter Ledger

3 free parameters · 3 axioms · 0 invented entities

The central null claims rest on the Poisson GLM framework, the independence of event counts, the chosen functional forms for year and report-threshold dependence, and the reliability of the AMS website tables (versus the inconsistent API). No new physical entities are introduced; free parameters are ordinary GLM coefficients fitted to the counts.

free parameters (3)
  • β_yr (year slope, log-link model) = 1.02 ± 0.05 (then set to 1)
    Fitted coefficient of ln(year–2009); later fixed to 1 after consistency check; controls the long-term growth rate used for all residual tests.
  • β_rep (report-threshold exponent) = −1.12 ± 0.04
    Power-law index of report-count bins; determines relative rates across size thresholds and is tested for 2026 interaction.
  • β_Qj / β_mk (quarterly and monthly intercepts) = see Tables 3–6
    Categorical offsets that absorb seasonal variation; used to construct expected rates against which 2026 residuals are judged.
axioms (3)
  • domain assumption Reported fireball counts in disjoint time and report-threshold bins are independent Poisson random variables whose means are functions of year, season, and threshold.
    Stated in §2.1 and used for all GLMs and residual diagnostics; independence is checked only for near-simultaneous events.
  • domain assumption Sun-centered ecliptic coordinates remove nodal-precession effects and are the appropriate frame for comparing radiant distributions across years.
    Justified in §2.4 and Fig. 3; used for the two-sample KS test in §3.5.
  • standard math Bonferroni correction with n_test equal to the number of residual cells (or tests) controls the family-wise error rate at α = 0.05.
    Applied in §2.3 and used for outlier thresholds in Figs. 6 and 8.

pith-pipeline@v1.1.0-grok45 · 19761 in / 2615 out tokens · 30263 ms · 2026-07-11T09:14:34.640519+00:00 · methodology

0 comments
read the original abstract

In March 2026, the American Meteor Society announced that a "surge" of large fireballs had been reported to their website in the first quarter of the year, and that these fireballs had certain characteristics (radiant clustering and reports of delayed sound). We find this data set to be an excellent use case for Poisson regression, which, in our opinion, is underutilized in meteor astronomy. This report serves as a brief primer on Poisson regression and related statistical techniques as well as an analysis of AMS fireball counts. We find that the number of events reported in early 2026 is in line with the overall pattern of activity. We also find little evidence of the "February fireballs" phenomenon.

Figures

Figures reproduced from arXiv: 2607.05071 by Althea V. Moorhead.

Figure 1
Figure 1. Figure 1: Example summary from a GLM run in R, with the most useful lines highlighted. necker, 2016). The rate can vary with time, location, or other external factors, but the expected number of events within an interval is the integral of the rate over that interval. It is not always necessary to use Poisson regression to ana￾lyze count data; when the expected count is large, a Poisson distribution resembles a norm… view at source ↗
Figure 2
Figure 2. Figure 2: In this diagram, the Earth’s orbit (dashed black circle) is intersected by three meteoroid orbits (gray ellipses) whose orbital elements are the same except for the longitude of ascending node. The Sun-Earth-radiant angle (thick red lines) – and thus the SCE radiant – is the same in all three cases. 0 ∘ 45 ∘ 90 ∘ 135 ∘ 180 ∘ 225 ∘ 270 ∘ 315 ∘ 360 ∘ solar longitude 0 ∘ 45 ∘ 90 ∘ 135 ∘ 180 ∘ 225 ∘ 270 ∘ 315 … view at source ↗
Figure 3
Figure 3. Figure 3: Radiants of meteors observed by the NASA All Sky Fireball Network in both equatorial (top) and Sun-centered ecliptic (SCE; bottom) coordinates. Points are color-coded by solar longitude (i.e., time of year). The position of the Sun is marked in the bottom panel. These data are not part of our analysis and are included here only to illustrate the utility of SCE coordinates. and center the data on an SCE lon… view at source ↗
Figure 4
Figure 4. Figure 4: shows the best fit to each subset of data. We no￾ticed some commonalities between these fits; for instance, all fits have similar x-intercepts (see [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: The number of fireballs reported to the AMS every year since 2011: each row corresponds to a different range in the number of reports submitted per fireball, and each column corresponds to a different quarter of the year. The solid line follows our best fit to eq. 9, and the shaded region approximates the 95% prediction interval [PITH_FULL_IMAGE:figures/full_fig_p007_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Quantile residuals for the number of fireballs reported to the AMS in each quarter year between the first quarter of 2011 and the first quarter of 2026 (inclusive). Residuals are calculated relative to the best fit to eq. 9 where βyr is set to 1. The dashed black lines at ±3.99 correspond to a Bonferroni-adjusted significance threshold of αtest = 0.5/244. There are no outliers present. ztest = 3.71, howeve… view at source ↗
Figure 7
Figure 7. Figure 7: Month-specific fit coefficient (βmk , black points) compared to the overall monthly average (horizontal dashed line). Error bars encompass the 95% confidence interval for each month. estimate std. error p-value βJan 3.04 0.16 1.5e−083 βFeb 3.17 0.15 3.0e−093 βMar 3.26 0.15 4.1e−100 βApr 2.78 0.17 8.3e−063 βMay 2.60 0.17 5.3e−052 βJun 2.70 0.17 2.0e−058 βJul 2.92 0.16 5.7e−073 βAug 2.90 0.16 1.9e−071 βSep 3… view at source ↗
Figure 8
Figure 8. Figure 8: Quantile residuals for the number of fireballs reported to the AMS in each month, year, and reporting interval. Residuals are calculated relative to the best fit to eq. 16. The dashed black lines at ±3.98 correspond to a Bonferroni-adjusted significance threshold of αtest = 0.5/732. There are no outliers present. words, the data do not show an anomalously high (or low) number of reported fireballs in any m… view at source ↗
Figure 9
Figure 9. Figure 9: Radiants of fireballs reported to the AMS, separated by year as labeled and shown in both equatorial (left) and Sun-centered ecliptic (SCE; right) coordinates. 4 Conclusions We were able to test five out of seven claims made by the American Meteor Society regarding reported fireball events in the first quarter of 2026. Preserving the numbering used in the introduction of this paper, our findings are as fol… view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

18 extracted references · 6 canonical work pages

  1. [1]

    Journal of Computational and Graphical Statistics 5, 236–244

    Randomized quantile residuals. Journal of Computational and Graphical Statistics 5, 236–244. doi:10.2307/1390802. Fasano, G., Franceschini, A.,

    doi:10.2307/1390802

  2. [2]

    Monthly Notices of the Royal Astronomical Society 225, 155–170

    A multidimensional version of the Kolmogorov–Smirnov test. Monthly Notices of the Royal Astronomical Society 225, 155–170. doi:10.1093/mnras/ 225.1.155. Fisher, R.A.,

    doi:10.1093/mnras/

  3. [3]

    Journal of the Royal Statistical Society 85, 87–94

    On the interpretation ofχ 2 from contingency tables, and the calculation of P. Journal of the Royal Statistical Society 85, 87–94. doi:10.2307/2340521. Gural, P .S.,

    doi:10.2307/2340521

  4. [4]

    Meteoritics and Planetary Science 31, 185–217

    Detailed data for 259 fireballs from the Canadian camera network and infer- ences concerning the influx of large meteoroids. Meteoritics and Planetary Science 31, 185–217. doi:10.1111/j.1945- 5100.1996.tb02014.x. Hankey, M.,

    doi:10.1111/j.1945- 1945

  5. [5]

    American Meteor Society fireball reporting system and mobile application, in: Asteroids, Comets, Meteors 2014, p

    2014

  6. [6]

    Has something changed in the near-Earth me- teoroid environment? URL: https://web.archive.org/web/ 20260326023100/https://amsmeteors.org/ams-q1-2026- fireball-analysis.html

    Hankey, M., 2026a. Has something changed in the near-Earth me- teoroid environment? URL: https://web.archive.org/web/ 20260326023100/https://amsmeteors.org/ams-q1-2026- fireball-analysis.html. Hankey, M., 2026b. Has something changed in the near- Earth meteoroid environment? (updated March 31, 2026). URL: https://web.archive.org/web/20260415194603/https: ...

    arXiv 2026

  7. [7]

    Springer International Publishing

    An Intro- duction to Statistical Learning: with Applications in R. Springer International Publishing. URL: https://www.statlearning.com/, doi:10.1007/978-1-0716-1418-1. James, G., Witten, D., Hastie, T ., Tibshirani, R., Taylor, J.,

    doi:10.1007/978-1-0716-1418-1

  8. [8]

    Springer International Publishing

    An Introduction to Statistical Learning: with Applications in Python. Springer International Publishing. URL: https:// www.statlearning.com/, doi:10.1007/978-3-031-38747-0. Johnson, N.L., Kemp, A.W ., Kotz, S.,

    work page doi:10.1007/978-3-031-38747-0

  9. [9]

    Univariate Discrete Distributions. Wiley . doi:10.1002/0471715816.ch4. Kingery, A., Moser, D.E., Cooke, W .J., Moorhead, A.V .,

    work page doi:10.1002/0471715816.ch4

  10. [10]

    doi:10.1051/0004-6361/ 202348618

    Astron- omy and Astrophysics 683, A5. doi:10.1051/0004-6361/ 202348618. LaPaz, L.,

    doi:10.1051/0004-6361/

  11. [11]

    Miaou, S.P ., Hu, P .S., Wright, T ., Rathi, A.K., Davis, S.C.,

    doi:10.1086/126127. Miaou, S.P ., Hu, P .S., Wright, T ., Rathi, A.K., Davis, S.C.,

    work page doi:10.1086/126127

  12. [12]

    Moser, D.E.,

    doi:10.3847/1538-3881/ae47e5. Moser, D.E.,

    work page doi:10.3847/1538-3881/ae47e5

  13. [13]

    Planetary Space Science 143, 182–191

    Comparing eyewitness-derived trajectories of bright meteors to instrumentally-observed data. Planetary Space Science 143, 182–191. doi:10.1016/j.pss.2017.02.016. O’Hara, R.B., Kotze, D.J.,

    work page doi:10.1016/j.pss.2017.02.016 2017

  14. [14]

    Methods in Ecology and Evolution 1, 118–122

    Do not log-transform count data. Methods in Ecology and Evolution 1, 118–122. doi:10.1111/ j.2041-210x.2010.00021.x. Ott, R.L., Longnecker, M.,

    Pith/arXiv arXiv 2041

  15. [15]

    Icarus 408, 115843

    GOES GLM, biased bolides, and debi- ased distributions. Icarus 408, 115843. doi:10.1016/ j.icarus.2023.115843. Phillips, T .,

    arXiv 2023

  16. [16]

    Phys.org URL: https: //web.archive.org/web/20190629031735/https://phys.org/ news/2012-02-fireballs-february.html

    The fireballs of February. Phys.org URL: https: //web.archive.org/web/20190629031735/https://phys.org/ news/2012-02-fireballs-february.html. Rendtel, J., Knöfel, A.,

    arXiv 2012

  17. [17]

    Space.com URL: https://web.archive.org/web/ 20120223225355/https://www.space.com/14663-nasa- february-fireballs-meteors-meteorites.html

    Strange fireballs light up February skies. Space.com URL: https://web.archive.org/web/ 20120223225355/https://www.space.com/14663-nasa- february-fireballs-meteors-meteorites.html. accessed: 2026- 04-13. Ver Hoef, J.M., Boveng, P .L.,

    2026

  18. [18]

    negative binomial regression: How should we model overdispersed count data? Ecology 88, 2766–2772

    Quasi-Poisson vs. negative binomial regression: How should we model overdispersed count data? Ecology 88, 2766–2772. doi:10.1890/07-0043.1. 11

    work page doi:10.1890/07-0043.1