---
analysis-role: frame-sequence
analysis-category: modern-military-control
confidence-level: medium
confidence-rationale: The paper identifies the exact local video and the raw sampled sequence confirms a coherent terrain-relative transit, but range, sensor intrinsics, platform telemetry, and the authors' full stabilization pipeline remain independently unverified.
evidence-status: mixed
ai-analysis: true
accuracy-disclaimer: AI-assisted analysis; interpretations are provisional and may contain errors. Verify against the cited video, paper, and any original sensor or platform records.
ai-generated: true
companion-eligible: true
topic-slug: pr043-djibouti-hypersonic-missile-reconstruction
topic-title: PR043 Djibouti Hypersonic-Missile Reconstruction
topic-description: Exact source pairing and raw-frame review of the Djibouti PR043 video against the Nandal-Loeb terrain, kinematic, and missile-physics reconstruction.
guide-description: Tests how strongly the new Galileo Project paper supports a fast tactical-missile interpretation without turning an altitude-dependent reconstruction into a unique speed or weapon identification.
topic-audience: serious-uap-readers
topic-priority: 9
---

# PR043 Djibouti Hypersonic-Missile Reconstruction

## Source Basis

This is an exact, high-triage source match. Devesh Nandal and Abraham Loeb's August 6, 2026 paper, [PR043 Djibouti Video Evidence](/?open=Release_1%2FAnalysis%2FPR043%20Djibouti%20Video%20Evidence.pdf), names the analyzed file as `DOD 111689759.mp4`. The local archive contains that exact source as [Unresolved UAP Report, Djibouti, 2025](/?open=Release_1%2FDOD_111689759.mp4).

The local file independently confirms the paper's basic acquisition facts: 1920x1080 pixels, 29.97 frames per second, 352 decoded frames, and 11.745 seconds of distributed duration. The paper reports that the short event is repeated about four times and analyzes one 88-frame, 2.94-second cycle. The total local frame count is exactly four 88-frame slots, so repeated loops must not be counted as four independent observations.

The paper is an authored Galileo Project analysis, not an agency determination. Its exact-file citation makes it a valid analysis companion to the government-released video; its conclusions remain secondary to the source pixels and conditional on the stated camera/geolocation model.

![Five raw full-frame samples from the first repeated cycle](/media/Release_1/Analysis/images/dod-111689759-pr043-track-roi-source-frame-contact-sheet.png)

The focused raw crop below covers source pixels `x=1008-1487`, `y=810-1079`. It samples the paper's stated frames 61-73 at [2.035s](/?open=Release_1%2FDOD_111689759.mp4&t=2.035), [2.135s](/?open=Release_1%2FDOD_111689759.mp4&t=2.135), [2.235s](/?open=Release_1%2FDOD_111689759.mp4&t=2.235), [2.336s](/?open=Release_1%2FDOD_111689759.mp4&t=2.336), and [2.436s](/?open=Release_1%2FDOD_111689759.mp4&t=2.436). It is a raw geometry aid, not an enhanced image.

![Raw PR043 track-region comparison](/media/Release_1/Analysis/images/dod-111689759-pr043-track-roi-geometry-review.png)

## Observation

The raw crop preserves a faint, compact contrast feature moving coherently from the upper-left toward the lower-right across the five samples. That direction and region agree with the 13-centroid path plotted in figure 2 of [PR043 Djibouti Video Evidence](/?open=Release_1%2FAnalysis%2FPR043%20Djibouti%20Video%20Evidence.pdf&page=3). The local pass therefore confirms that the reported track refers to a source-visible moving feature rather than to a single selected still.

This pass does not independently reproduce the paper's full terrain registration, projective transform, Monte Carlo perturbation, camera fit, or acceleration bootstrap. Those are evaluated here as reported methods and results:

| Question | Source-backed result | Boundary |
| --- | --- | --- |
| Is there a moving feature? | Yes. The raw first-loop samples show a small contrast feature crossing the terrain region over about 0.40 second. | The target is unresolved and only a few pixels wide. |
| What is directly measured? | The paper reports a terrain-relative image speed of about `801.3 px/s` after time-dependent homography correction. | Pixels per second are not metres per second. |
| Where is the terrain? | The paper's best match is near `11.388655 N, 43.139259 E`, close to Goubetto in southern Djibouti. | The coordinate is a ground anchor, not the sensor or target position. |
| What does a ground-plane intersection yield? | A central result near `3.33 km/s`, with a Monte Carlo median of `3.37 km/s` and a reported 95% interval of `2.52-4.72 km/s`. | The target track lies outside the strongest control-point convex hull, so this is extrapolated and conditional. |
| What happens as target altitude rises? | The paper reports about `3.54 km/s` at 0 m, `2.74 km/s` at 500 m, `1.94 km/s` at 1,000 m, `1.14 km/s` at 1,500 m, and `0.34 km/s` at 2,000 m. | The video provides no defensible target-altitude prior, so these values form a family, not a probability distribution. |
| Is acceleration detected? | No selected altitude shows statistically significant acceleration or deceleration in the paper's 0.4004-second window. | Constant apparent speed cannot distinguish powered flight from a dense body coasting over such a short interval. |
| Is missile scale supported? | The paper derives a blurred cross-track upper limit of roughly 4-9 m for assumed altitudes of 0.5-1.5 km and finds a fast tactical missile or brief terminal passage physically compatible with that branch. | The width includes point-spread, compression, motion blur, and thermal blooming; it is not a measured missile length or type. |

## Hypothesis To Test

The primary hypothesis is that PR043 records a fast tactical missile or terminal missile passage rather than an anomalous vehicle. The paper improves this from a filename-level guess to a physics-constrained control model: speed, Mach number, dynamic pressure, heating severity, ballistic coefficient, drag, and apparent width must all be mutually compatible at the same assumed altitude.

Three explanation lanes remain live:

1. **Fast tactical missile or terminal passage.** This is the paper's strongest interpretation. It is compatible with part of the altitude-dependent solution family, and the short sequence is too brief to require detectable powered acceleration.
2. **Closer conventional object or image-chain effect.** Near the conditional sensor altitude, the inferred speed and size fall sharply. A conventional aircraft, drone, bird, or optical/infrared effect becomes more plausible only in this closer, smaller, lower-speed regime.
3. **Advanced or disclosure-forward vehicle.** A compact vehicle using unconventional propulsion is not ruled out by the pixels alone, especially if later evidence fixes the target at the distant/high-speed branch while excluding a missile. The current video does not supply that range, nor does it show a resolved hull, field effect, environmental response, or non-ballistic maneuver that positively supports this lane.

## Theoretical Scene Panels

| Panel | Source-grounded read | Interpretive model | Limit |
| --- | --- | --- | --- |
| A - Terrain crossing | The feature moves diagonally across the stabilized terrain region from [2.035s](/?open=Release_1%2FDOD_111689759.mp4&t=2.035) to [2.436s](/?open=Release_1%2FDOD_111689759.mp4&t=2.436). | A compact fast target crosses the line of sight while the sensor platform also moves. | Image-plane motion alone does not give range or true speed. |
| B - Range fork | The same angular track maps to kilometres per second near the terrain and a few hundred metres per second near the modeled sensor altitude. | The event is a range-degenerate kinematic fork, not a single-speed measurement. | No target-altitude prior is available. |
| C - Missile physics | The low-altitude branch carries severe dynamic pressure and heating; a dense, streamlined body can still lose only a few percent of speed in 0.4004 second. | A brief terminal missile passage fits better than sustained low-altitude hypersonic cruise. | No plume, launch, impact, or weapon-system identity is visible. |
| D - Exotic falsification lane | An advanced-vehicle interpretation survives only as an unresolved alternative. | It would strengthen if independent range fixed the target on the distant branch and weapons/context checks failed. | It weakens if platform logs, launch timing, plume evidence, or trajectory identify a conventional missile. |

## Modern Military Tie-In

PR043 belongs in [C03 - Modern Military Operational Cases](/?open=Release_2%2FAnalysis%2FC03-Modern-Military-Operational-Cases.md) and [C14 - Conventional, Misleading, and Balloon Flags](/?open=Release_2%2FAnalysis%2FC14-Conventional-Misleading-and-Balloon-Flags.md) as a high-value conventional-control case. Its value is not that it proves a named missile. Its value is that an exact-source scientific reconstruction turns a fast-dot video into explicit, falsifiable constraints.

Do not merge PR043 with [Mission Report, Djibouti, 2025](/?open=Release_1%2FDOW-UAP-D8-Mission-Report-Djibouti-2025.pdf) merely because both carry a Djibouti label. D8 describes two round white-hot UAPs moving south at an estimated speed; PR043's official metadata describes one barely distinguishable contrast feature crossing the frame. Without a shared event identifier, date/time, platform record, or explicit citation, D8 remains a contextual Djibouti lead rather than a factual companion to this video.

## Why It Matters

This paper materially changes the review posture for [Unresolved UAP Report, Djibouti, 2025](/?open=Release_1%2FDOD_111689759.mp4). The source is no longer just a faint rapid transit. It now has a specific geolocation proposal, a terrain-registered image-speed measurement, a conditional altitude-speed-size family, and an aerodynamic test of the missile explanation.

The strongest result is also the narrowest: a real source-visible feature moves rapidly relative to the terrain, but the missing range prevents a unique physical speed. The missile interpretation is the best-supported current control hypothesis, not an identification certificate.

## Working Assessment

Promote [C90 - PR043 Djibouti Hypersonic-Missile Reconstruction](/?open=Release_1%2FAnalysis%2FC90-PR043-Djibouti-Hypersonic-Missile-Reconstruction.md) as the exact companion for both the PR043 video and the Nandal-Loeb paper. Assign high triage for review and medium confidence in the working interpretation.

The evidence supports a fast tactical missile or brief terminal missile passage as the strongest present explanation. It does not determine target altitude, unique speed, propulsion state, or missile type. The paper's low-altitude hypersonic branch is aerodynamically demanding, while the closer/high-altitude branch lowers speed and apparent size; neither branch is selected by the released video alone.

The Disclosure-Forward Neutrality Gate passes. Conventional missile, aircraft, close-object, and image-chain controls lead because they fit the current source without requiring new physics. The advanced-vehicle lane remains testable rather than dismissed, but it would need independent range, calibrated sensor/platform data, non-ballistic maneuver, multi-sensor persistence, or physical effects that defeat the missile family. A launch/weapon-log match would weaken or falsify the exotic lane.

## Follow-Up

- Reproduce the paper's homography from published control points and report the target-region extrapolation sensitivity separately from the ground-control fit.
- Obtain original sensor intrinsics, platform position/velocity, slant range, and native non-looped frames before converting `801.3 px/s` into one physical speed.
- Check mission and weapons logs for a launch corridor, launch time, missile family, impact, or terminal-pass context consistent with the reconstructed bearing near `65.3 degrees true`.
- Test for a plume or wake across native frames without enhancement-driven promotion; the present public clip does not show a clear one.
- Keep D8 as a separate contextual lead unless an exact identifier or event-level source explicitly joins it to PR043.