🔭 astrophysics

Independent Replication of Nuclear Test-Transient Correlations and Earth Shadow Deficit in POSS-I Photographic Plates

This paper presents an independent replication of findings by Bruehl and Villarroel (2025) and Villarroel et al. (2025), confirming a statistically significant correlation between transient sources on POSS-I photographic plates and atmospheric nuclear weapons tests, as well as a deficit of such sources within Earth's geosynchronous shadow, using the original dataset and advanced statistical controls.

Original authors: Brian Doherty

Published 2026-04-02

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Brian Doherty

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are looking at a giant, old-fashioned photo album from the 1950s. These aren't photos of your family; they are photos of the night sky taken by a telescope at Palomar Observatory. For decades, astronomers thought these photos only showed stars and planets that stayed in the same place.

But recently, a team of researchers noticed something weird: there are thousands of "ghosts" in these photos. These are tiny dots of light that appear on just one photo and then vanish. They aren't stars, they aren't planes, and they weren't satellites (because the first satellite, Sputnik, hadn't been launched yet).

A new paper by an independent researcher named Brian Doherty is like a "second opinion" from a different doctor. He took the same data the original researchers used and ran his own tests to see if their findings were real or just a fluke.

Here is what he found, explained simply:

1. The "Nuclear Fireworks" Connection

The Original Claim: The researchers noticed that on days when the US government tested nuclear bombs in the atmosphere, the telescope photos were much more likely to show these mysterious "ghost" dots. It was like the sky was lighting up with extra sparks whenever a bomb went off.

The Replication: Doherty checked the math. He compared days with nuclear tests to days without them.

The Result: He confirmed the original finding. On days near a nuclear test, the chance of seeing these ghosts jumped by about 45%.
The Analogy: Imagine you are counting how many times a specific type of bird flies over your house. You notice that every time a construction crew starts drilling nearby, the birds appear more often. Doherty checked the logs and said, "Yes, the drilling definitely correlates with the birds showing up."

2. The "Shadow Game"

The Original Claim: The researchers also noticed that these "ghosts" almost never appear in a specific part of the sky: the shadow of the Earth.
Think of the Earth casting a giant, invisible cone of darkness into space (like a flashlight beam pointing away from the sun). If these ghosts were just random space junk or satellites reflecting sunlight, they should appear everywhere, including inside that shadow. But they don't. They seem to avoid the shadow like moths avoiding a dark closet.

The Replication: Doherty mapped every single ghost against the Earth's shadow.

The Result: He confirmed the "avoidance." Only about 0.45% of the ghosts were found in the shadow, whereas geometry says there should be about 1.4%.
The Analogy: Imagine you are looking for raindrops on a window. If the rain is random, some drops should be on the part of the window covered by a tree's shadow. But if you find that zero drops are under the tree, you might guess that the "rain" isn't actually rain—it's something that needs the sun to be visible.

3. The "Sunlight" Clue

This is the most interesting part. Doherty ran a special test: he looked only at the ghosts that were in the sunlight (not in the shadow).

The Result: When he looked only at the sunlit ghosts, the connection to the nuclear tests got much stronger (nearly 4 times stronger).
The Analogy: Think of it like this: If you are trying to see a shiny coin in a dark room, you need a flashlight. If the "ghosts" are shiny objects reflecting sunlight, they would only be visible when the sun hits them. The fact that the nuclear test connection is strongest when the objects are in the sun suggests they might be reflecting light, not glowing on their own.

4. Why This Matters (and Why It's Weird)

Doherty's job was to make sure the original researchers didn't make a mistake. He used different computers, different math, and different ways of checking the data.

The Verdict: The original findings are statistically solid. The connection between nuclear tests and these sky ghosts is real, and the avoidance of Earth's shadow is real.

The Big Mystery:
The paper doesn't say what these objects are. It just confirms that they exist and behave strangely.

They appeared before satellites existed.
They seem to reflect sunlight (because they hide in the shadow).
They appear more often when nuclear bombs are tested.

The "So What?"
If these are natural phenomena (like stars or asteroids), they wouldn't care about nuclear tests or Earth's shadow. If they are man-made debris, there were no man-made satellites in space in the 1950s to create them.

Doherty's paper is like a detective saying, "I've double-checked the evidence. The clues are real. We know when they happen and where they hide, but we still don't know who or what they are."

Summary in One Sentence

An independent researcher confirmed that mysterious, fleeting lights in 1950s sky photos appear more often during nuclear tests and avoid Earth's shadow, suggesting they might be reflective objects that require sunlight to be seen, though their true identity remains a complete mystery.

1. Problem Statement

The paper addresses two striking claims made by Bruehl and Villarroel (2025) and Villarroel et al. (2025) regarding the VASCO (Vanishing and Appearing Sources during a Century of Observations) catalog of transient sources on Palomar Observatory Sky Survey (POSS-I) photographic plates (1949–1958).

The core hypotheses being tested are:

Temporal Correlation: There is a statistically significant increase in transient detection rates on days within $\pm 1$ day of atmospheric nuclear weapons tests compared to non-test days.
Spatial Deficit: There is a significant deficit of transient sources located within Earth's geometric shadow cone (umbra) at geosynchronous orbit (GEO) altitude.

The significance of these claims lies in the fact that all transients predate the launch of Sputnik 1 (October 1957). If the transients are reflective objects in orbit, their correlation with nuclear tests and avoidance of Earth's shadow suggests a physical mechanism distinct from known astrophysical phenomena or pre-satellite era artifacts. Independent replication is required to rule out statistical artifacts, temporal autocorrelation, or environmental confounding before investing in physical interpretations.

2. Methodology

The author performed an independent statistical analysis using the original dataset provided by the original authors (107,875 transients across 635 plates). The analysis was conducted using code written independently of the original authors' implementations.

Key Methodological Components:

Data Sources:
- Transient Catalog: 107,875 transients detected by the VASCO pipeline.
- Nuclear Test Data: Official US DOE/NV-209 Rev 16 dates, defining a test window as $\pm 1$ day of detonation.
- Environmental Covariates: Moon illumination (Astropy), cloud cover (NOAA ISD, San Diego station), and precipitation (NOAA GHCND).
Statistical Tests:
1. Chi-Square Contingency Test: Compared transient detection rates inside vs. outside nuclear test windows to calculate Relative Risk (RR).
2. Negative Binomial Regression: Modeled daily transient counts to control for overdispersion and environmental confounders (moon phase, clouds, rain, and UFOCAT sighting counts). The model tested the Incidence Rate Ratio (IRR) for the nuclear window variable.
3. Permutation Test: A 10,000-iteration shuffle of nuclear test labels to determine if the correlation was driven by specific dates or general temporal autocorrelation. Block permutations (30, 60, 90 days) were also used to preserve temporal structure.
4. Earth Shadow Classification: Calculated the angular separation between each transient and the anti-solar point at GEO altitude. Transients within an $8.50^\circ$ radius were classified as "in shadow."
Robustness Checks:
- Center-of-Plate Subset: Restricted analysis to sources within $2^\circ$ of the plate center to eliminate edge artifacts.
- Sunlit vs. Shadowed: Separated transients based on whether they were in Earth's shadow to test the "solar reflection" hypothesis.
- Model Selection: Compared Poisson, Negative Binomial, and Zero-Inflated models using AIC.

3. Key Contributions

Independent Verification: Confirmed the core statistical findings of the original papers using an independent codebase and dataset.
Confounder Control: Extended the original analysis by rigorously controlling for environmental variables (clouds, precipitation, moon phase) and UFO sightings, demonstrating the nuclear correlation persists.
Temporal Specificity: Proved via permutation tests that the correlation is driven specifically by the dates of nuclear detonations, not by general seasonal clustering or temporal autocorrelation in the data.
Physical Consistency Check: Demonstrated that the nuclear correlation is strongest for transients in sunlit positions (IRR $\approx$ 4.0) and weaker for all transients, supporting the hypothesis that the objects require sunlight to be visible (consistent with reflective objects).
Shadow Deficit Quantification: Confirmed the deficit of transients in Earth's shadow is robust to plate-edge filtering and non-uniform sky coverage.

4. Key Results

A. Nuclear Test Correlation

Chi-Square: Transient detection rate was 15.6% during test windows vs. 10.8% outside, yielding a Relative Risk (RR) of 1.45 ( $p = 0.011$ ).
Regression Analysis:
- All Transients: The nuclear window IRR was 1.80 ( $p < 0.0001$ ) after controlling for covariates.
- Sunlit-Only Transients: The IRR increased to 3.98 ( $p < 0.0001$ ), nearly four times the expected count during test windows. This aligns with the original paper's findings (IRR $\approx$ 3.53).
Permutation Test: The observed RR of 1.45 exceeded 99.4% of the permuted distribution ( $p = 0.006$ ). Block permutation tests confirmed the result is not an artifact of temporal autocorrelation.

B. Earth Shadow Deficit

Observed Rates: Only 0.46% (499/107,875) of the full catalog and 0.45% (142/31,525) of the center-of-plate subset fell within the GEO shadow cone.
Expected Rates: The geometric expectation for uniform sky coverage is $\approx 1.4\%$ . Even after correcting for the non-uniform plate coverage (which lowers the expectation to 0.78%), the observed rates are significantly lower.
Conclusion: The deficit is statistically significant and robust to edge artifacts.

C. Robustness

The results held across different time windows ( $\pm 2$ and $\pm 4$ days), different time periods (early vs. late survey), and various distributional assumptions (Poisson, NB1, NB2, Zero-Inflated).

5. Significance and Implications

Statistical Robustness: The study confirms that the association between nuclear tests and POSS-I transients is statistically robust, surviving rigorous controls for environmental factors and temporal structure.
Physical Constraints: The findings impose strict constraints on any physical explanation for these transients. A viable hypothesis must explain:
1. Why the appearance rate correlates with nuclear detonations.
2. Why the objects avoid Earth's shadow (implying they are reflective and require sunlight).
3. Why they appear before the existence of artificial satellites.
Exclusion of Alternatives: Conventional astrophysical transients (variable stars, asteroids) would not exhibit this specific temporal correlation or spatial shadow avoidance. Atmospheric phenomena or plate artifacts might correlate with time but would not respect the precise geometric shadow of a specific orbital altitude.
Future Directions: While the statistical correlation is confirmed, the paper explicitly states that causation is not established. The physical mechanism remains unknown. The results suggest the need for cross-validation with other observatories (e.g., Hamburg Observatory) and further investigation into whether these transients represent a previously unclassified population of reflective objects.

Conclusion: The paper successfully replicates the original findings, confirming that POSS-I transients show a significant, non-random correlation with nuclear test dates and a significant deficit in Earth's shadow, challenging conventional explanations for the nature of these pre-satellite era anomalies.