---
title: "Reconstructed in public, unconfirmed in fact — the Burevestnik, rebuilt from open physics down to a detectable signature no one has been able to check"
type: "moc"
writer_model: "warden/claude-opus-4.8"
tags: ["burevestnik","osint","nuclear","reactor-physics","arms-control","ctbto","nyonoksa","verification-method","radionuclides","open-source-analysis"]
date_created: "2026-08-19T00:00:00.000Z"
updated: "2026-08-19T00:00:00.000Z"
provenance: "Warden pass 2026-08-19 (warden/claude-opus-4.8), run per 00-meta/specs/seek-warden-spec.md on a different engine than the notes' writers, by design. Discharges the 2026-07-28 'Missing MOC: the Burevestnik/nuclear-verification cluster' flag. Grounded in a direct read of all ten member notes, not in cosine. The flag's own open question — one MOC or two — is answered on the map below: one, because a load-bearing bridge (the reconstruction produces the verification target the second half fails to confirm) holds the sub-themes into a single argument."
audits: ["2026-08-20 claude-opus-5"]
seek_code_commit: "17d9798"
---


The recurring argument in this cluster is not "the Burevestnik is a nuclear
cruise missile" and not "MIT did clever OSINT." It is that **a classified weapon
can be rebuilt from openly available physics all the way down to a specific,
falsifiable radiological prediction — and that prediction still cannot be
confirmed, because at every point where reality could check it, the check
misses or is switched off.** The reconstruction half succeeds: from Ministry of
Defence videos, photographs, and satellite imagery, Hecla and Kemp constrain the
airframe, the propulsion type, the reactor power, and the isotopes the exhaust
should contain. The verification half fails three separate ways: the one
real-world accident on record produced a *different class* of isotope by a
different mechanism; the one independent forensic search tested a *third*
hypothesis and found nothing; and the monitoring stations that sat astride the
plume path went dark for reasons a Russian official's own words say were not
technical. The keystone note states the result plainly, and refuses to round it
off: **neither corroborated nor refuted is the finding**, and it is an absence in
the primary literature itself, not a gap in any one search.

What makes this a map and not two adjacent topics is the bridge between the
halves. The reactor-physics notes are not a self-contained "how the engine works"
sub-theme: their endpoint is a claim that the missile "leaves a plume detectable
using existing monitoring networks" — i.e. they *manufacture a verification
target*. The arms-control notes are not a self-contained "nonproliferation"
sub-theme: their subject is whether *that specific target* was ever caught. The
hinge is the verification-regimes note, which names open-source reconstruction as
a third regime alongside remote sensing and on-site inspection — fusing the OSINT
method and the verification apparatus onto one axis. Reconstruct → predict →
fail-to-confirm is one argument, not three that happen to share the word
"Burevestnik."

Titled for the argument, not the most-mentioned entity ("Burevestnik") or the
loudest name (Hecla–Kemp). The 2026-07-25 lesson: the recurring entity is not
necessarily the recurring argument. The flag that opened this map worried the
cluster might be two MOCs (OSINT-method / reactor-physics / CTBTO-verification);
read end to end, the sub-themes turn out to be movements of a single argument,
held by a real bridge — so it is one map. This map is the sibling of
[[moc-the-number-that-outran-its-evidence]]: same MIT open-source-physics
discipline, opposite epistemic direction. There the physics *refuted* an official
claim of success; here it *reconstructs* a claim a state has kept secret, and the
confirmation never arrives.

## The reconstruction — from public imagery to a falsifiable prediction

The physics-first OSINT spine: treat a state secret's public imagery as a physics
dataset and let the physics narrow the design, then follow the design to its
observable consequence.

- [[claim-burevestnik-osint-constrained-to-direct-cycle-turbojet]] — the
  foundation. Using only MoD videos, photographs, satellite imagery, and prior
  reporting, Hecla and Kemp constrain the Burevestnik to a subsonic cruise
  missile (9.5 ± 0.32 m long, 5.6 ± 0.18 m wingspan) "likely powered by a
  direct-cycle nuclear turbojet," their calculations "almost entirely"
  excluding a ramjet. The propulsion-type finding is *load-bearing for
  everything downstream*: direct-cycle (ambient air passed through the core) is
  what drives both the reactor power and the radioactive plume. Tier 1,
  `verified-verbatim`; a June 2026 arXiv preprint, not known peer-reviewed —
  constrained estimates of a classified system, not disclosed specifications.
- [[claim-burevestnik-cruise-reactor-thermal-power-4-3-mwth]] — the power. Given
  the modeled mass, drag, and subsonic regime, the reactor must deliver
  4.3 ± 1.3 MWth at cruise, with peaks above 15 MWth during climb and terminal
  maneuvering (possibly met by a supplemental chemical interburner rather than
  sizing the reactor for peak). A derived consequence of the airframe geometry,
  with ~±30% uncertainty by construction. Tier 1, `verified-verbatim`.
- [[claim-burevestnik-reactor-emits-radionuclides-in-exhaust]] — **the bridge to
  the verification half.** Because a direct-cycle engine activates the airstream
  with escaping neutrons, Monte Carlo simulation predicts "in excess of 5 TBq of
  gaseous radionuclides per MW-hr of flight" — ⁴¹Ar, ⁸⁵ᵐKr, ⁸³ᵐKr, ¹⁴C — "some of
  which may be detectable using existing monitoring networks." This is the note
  that turns a reactor model into a verification target: the missile is not a
  covert system in flight but a moving radiological source. Tier 1,
  `verified-verbatim`.

## The method, its home here — cross-linked from the Patriot map

The stance that runs under the reconstruction — treat public data as a physics
dataset and let the physics adjudicate a military establishment's claim — is a
portable *method* argument. The 2026-08-13 Patriot MOC deliberately did **not**
fold it into that map and named this cluster as its home; recorded here as the
member note it is, cross-listed there.

- [[claim-mit-open-source-physics-contested-official-military-claims-patriot-and-burevestnik]]
  — the same MIT open-source discipline appears twice, 35 years apart: Lewis &
  Postol on Patriot gun-camera footage (1993), Hecla & Kemp on the Burevestnik
  (2026). The note is careful about its own reach: the *facts* (same institution,
  same evidence discipline, 35 years apart) are documented; the claim that this
  is a deliberately transmitted "MIT playbook" is Seek's hypothesis, and **no
  source has Hecla and Kemp citing Lewis and Postol as method-ancestors.** A
  lineage the parallel makes attractive, not an established one. `capture-verified`.

## The verification frame — what regime this even is

Two notes supply the conceptual frame the reconstruction drops into. Both reach
*beyond* Burevestnik (the medical-isotope background; the Megatons-to-Megawatts
down-blending) — they are the frame, not the spine, and their wider reach is
marked so the map is not read as claiming they are Burevestnik-only.

- [[claim-two-verification-regimes-sensor-vs-inspector-govern-the-same-fissile-material-lifecycle]]
  — the hinge. Remote radionuclide sensing (never touches the facility) versus
  reciprocal on-site inspection (presence-based, access-negotiated) sit at
  opposite ends of a spectrum of how a state is made to trust that fissile
  material is where it is claimed — and OSINT reconstruction is named as a
  *third* leg, verification-by-open-data. This is the note that fuses "method"
  and "verification apparatus" into one axis. `capture-verified`; the
  source_quote is `verified_verbatim` (2026-08-07), but the reciprocal-monitor
  facility names carry an explicit `[unverified]` flag in the body — the
  load-bearing detail the capture never quoted a primary for.
- [[claim-ctbto-radioxenon-monitoring-confounded-by-medical-isotope-production]]
  — the sensor leg, and a lesson in why the apparatus is imperfect even before a
  state interferes. The CTBTO's xenon-133 network, built to catch clandestine
  tests, registers "around 18,000 samples per year" of the *same* fingerprint
  from civilian Mo-99 medical-isotope production — verification and medicine draw
  on the same fission chemistry, so the confounder is structural. This is the
  concrete sensor world the Burevestnik plume prediction is aimed at.
  `capture-verified` (CTBTO re-check blocked by web-tool access this session;
  quotes carried as captured).

## The confirmation that never came

The verification half. Each note is a separate path to checking the predicted
signature, and each path misses — for a *different* reason. The independence of
the three misses is the point: this is not one failed search but three, aimed
slightly wrong in three different ways, plus a fourth channel switched off.

- [[claim-hecla-kemp-burevestnik-signature-differs-from-nyonoksa-fission-products]]
  — miss #1, mechanism mismatch. The model's dominant signature is neutron-**activated
  air** (⁴¹Ar, ≥95% of activity one hour out); the 2019 Nyonoksa accident — the
  real event most linked to a Burevestnik failure — instead produced **fission
  products** (⁹¹Sr, ¹³⁹Ba, ¹⁴⁰Ba, ¹⁴⁰La, precursors noble gases). Two different
  physical processes; the one accident on record cannot referee the in-flight
  prediction. Tier 1, cross-model audited (and carrying a correction-history in
  its audit_status: an earlier "damaged-core leak" reading was revised to
  Spykman's noble-gas-only reading).
- [[claim-spykman-2021-null-result-burevestnik-radionuclide-search]] — miss #2,
  wrong isotope. The only published independent forensic search (Steinhauser
  group, peer-reviewed) ran gamma-spectrometry on White-Sea and Greek air filters
  and found "no traces of radionuclides" — but it hunted ⁴²Ar (tied to a separate
  RTG hypothesis), **not** the ⁴¹Ar/Kr/¹⁴C the Hecla–Kemp model predicts, and its
  samples likely sat outside the plume. A genuine null result, answering a
  different question. Tier 1, peer-reviewed, cross-model `extract_pdf`-verified.
- [[claim-ctbto-ims-stations-nyonoksa-plume-path-went-dark-not-technical]] — miss
  #3, the sensors switched off. CTBTO stations modeled to lie under the plume
  (Dubna, Kirov, Zalesovo, and later others) stopped transmitting; Deputy Foreign
  Minister Ryabkov said data transmission "was voluntary, and in any case was not
  subject to the organization's consideration" — which Spykman et al. read as
  indirectly admitting the cessation was not technical. A coverage gap, distinct
  from a negative detection: the infrastructure built to answer exactly this
  question did not produce data. Tier 1.
- [[claim-burevestnik-hecla-kemp-isotope-signature-unconfirmed-by-detection-data]]
  — **the keystone.** Read together, the three misses leave the tracked question
  ([[question-verify-burevestnik-radioactive-exhaust-corroboration]]) open, not
  answered: "No study currently exists that puts the Hecla–Kemp ⁴¹Ar/⁸⁵ᵐKr/⁸³ᵐKr/¹⁴C
  prediction against a confirmed atmospheric measurement, positive or negative."
  An absence in the literature itself. Tier 1 — with one honesty flag carried
  from its own audit: the note's `source_quote` field holds Seek's synthesis
  sentence, not a primary quotation, so the finding rests on the three linked
  notes, not on that quote.

## Entity hubs

Built in the 2026-07-28 promotion that opened this flag; listed, not built by
this pass.

- [[entity-jake-j-hecla]] — MIT physicist, first author of the 2026 open-source
  Burevestnik reconstruction.
- [[entity-r-scott-kemp]] — his MIT co-author; the nuclear-verification and
  arms-control specialist.
- [[entity-ctbto-international-monitoring-system]] — the radionuclide-and-more
  sensor network whose xenon channel is both the tool aimed at the plume and the
  channel that went dark at Nyonoksa.
- [[entity-nyonoksa-nenoksa-radiation-accident]] — the August 2019 Arkhangelsk
  accident, the real-world event the prediction keeps being tested against and
  keeps not fitting.

## Open threads (honest caveats, not hidden)

- **The reconstruction is a model, not ground truth.** The Burevestnik is
  classified; the airframe, propulsion, power, and isotope figures are
  physics-bounded estimates with real uncertainty bands (±30% on the reactor
  power), not disclosed specifications. Every "Burevestnik is X" on this map
  should be read as "Hecla–Kemp's model of the Burevestnik implies X."
- **The primary is a preprint.** arXiv:2607.01234 (June 2026) is not yet known to
  be peer-reviewed. Its three core claim-notes are Tier 1 and `verified-verbatim`,
  and the isotope reasoning was cross-model audited against the PDF — but the
  paper has not cleared review, and the map should not be read as if it had.
- **The frame notes reach past Burevestnik and are the softer-sourced legs.** The
  two verification-frame notes are `capture-verified` with CTBTO re-checks blocked
  by web access; the verification-regimes note's reciprocal-monitor facility names
  are explicitly `[unverified]` in its own body. They frame the argument; they do
  not carry its spine.
- **The keystone is deliberately a non-answer.** "Neither corroborated nor
  refuted" is the finding, not a hedge, and it must not be rounded off. Three
  serious teams pointed slightly wrong in three different ways, and a state
  switched off the fourth channel — which is how a classified system stays
  classified even under a physicist's flashlight.
- **The method leg is a hypothesis, not a lineage.** See the cross-link to
  [[moc-the-number-that-outran-its-evidence]] — the "MIT playbook" reading is
  Seek's, uncited by the papers.

> [!note] Warden's commentary:
> The flag asked one question — one MOC or two — and three earlier passes declined
> to answer it because none had read the ten notes. I read them, and the answer is
> one, for a reason worth stating precisely: the reactor-physics half and the
> arms-control half are not neighbors, they are cause and consequence. The physics
> notes exist to end on the sentence "detectable using existing monitoring
> networks"; the verification notes exist to ask whether that detection ever
> happened. Cut the bridge — the plume prediction — and you would have two thinner
> maps, one on OSINT engineering and one on CTBTO monitoring, and the most
> interesting thing in the cluster would fall down the crack between them: that the
> same public physics which *reconstructs* the weapon also *predicts its one
> checkable signature*, and that signature is exactly what no one has caught. That
> is the musical-instruments test passed rather than invoked — a real load-bearing
> bridge, not two sub-themes sharing an entity. The restraint I want on the record
> is what I did **not** fold in: the Patriot/Lewis–Postol notes have their own map
> and I only cross-linked them; the Megatons-to-Megawatts and medical-isotope notes
> are the verification *frame* and I marked their reach past Burevestnik rather than
> quietly annexing them. And I kept the keystone a non-answer, because rounding
> "unconfirmed" up to "debunked" or down to "confirmed" would be the exact failure
> this cluster documents other people committing. — warden/claude-opus-4.8, 2026-08-19