Link to the paper I quoted from is right there under the graph, but I'll link again:
https://arxiv.org/pdf/2609.09461
This was a great post, Dave, and thanks for taking the time to cover the main points in a way that helps those of us who aren't familiar with the field or paper. Sorry, I didn't realise that was a link to the paper. I'm not familiar with that journal.
I initially misunderstood which earlier dataset was being used here. The paper states that the source is the much larger 107,875-transient dataset derived from Solano et al. (2022) and further detailed in later work. It is not the nine-transient April 1950 case from the 2021 paper. In Methods it states:
"The initial source dataset was a sample of n = 107,875 transients derived by Solano et al. [1] and further detailed in Villarroel et al. [9]. For each analysis described below, we then selected from this full dataset only transients determined in our prior work [10] to have a probability of ≥0.70 (excess analysis) or ≥0.90 (cluster analysis) of being a real transient rather than a plate defect."
The 2021 "nine simultaneously occurring transients" paper is separately listed as reference [3], whereas [1] is Solano et al. (2022). So the 203 Castle Yankee candidates are being selected from this much larger dataset, not from those nine 1950 objects.
So I agree it may be possible to work backwards from the date, POSS-I logs and coordinates to identify the plate(s), but that seems unnecessarily difficult when the analysis itself already knows which plate each transient came from.
I initially thought the "reasonable request" referred to access to physical plates, but on rereading it explicitly says the
dataset:
"The final dataset reflecting all analyzed variables will be made available by the authors upon reasonable request to Dr. Beatriz Villarroel."
I'm unsure what the convention is in this field, but as an outsider it seems odd not to provide the plate IDs when plate-level analysis was performed.
I also agree about the assumed altitude. The paper states that
"all projected locations depended on the assumed fixed altitude" of 35,786 km. I understand there is another paper Villarroel et al. preprint
Modelling Palomar Transients: Constraints from Reflection Geometry and Orbital Altitude (September 2026) that addresses this novel approach of establishing the altitude, but why not also show the projected locations at, say, 20,000, 35,786 and 50,000 km, simply to demonstrate how sensitive the claimed geographical hotspots are to that assumption?
The significance the authors attach to the locations also seems important here. In the Discussion they say:
"Nearly 90% of the highest probability transients associated with nuclear testing (n = 203) were observed exclusively over Pacific region hotspots identified in this work, all immediately prior to the second largest ever US thermonuclear detonation"
But this raises another question for me:
why should the phenomenon occur before the second-largest nuclear test rather than the largest, and why before rather than during or after? Their association is defined using a ±1-day window. For Castle Yankee all 203 happened the night before, whereas for the Nevada tests they report transients both on the night of a test and, notably, 23 on the night after one test.
The authors go considerably further when discussing the "before" result:
"one tentative interpretation of our findings could be that transients display characteristics suggesting possible intelligence, that is, a preference for specific locations over the Earth that may fit UAP lore and behavior suggesting possible awareness of imminent nuclear tests."
They immediately acknowledge:
"While this hypothesis may fit available data, it is unfortunately not falsifiable."
That seems important. If the proposed significance of the
preceding observations ultimately requires something to somehow anticipate an imminent test, I'd first want to understand whether there is a more ordinary reason for the temporal pattern and why the ±1-day window is the appropriate one.
As an outsider to the field, and someone more interested in what the papers actually establish than in the surrounding drama, I'm finding it increasingly difficult to assess this later work when the fundamental problem of possible plate defects still seems unresolved. More assumptions and unsupported hypotheses are being built on top of detections whose physical origin has yet to be firmly established. I may not look at these papers or anything surrounding them for a long while.
