The Crozon Peninsula Red Rectangles
Early on the morning of January 13, 2020, a driver on a small country road on the Crozon peninsula, Finistère, was driving through a storm — heavy wind gusts, a low cloud ceiling, rain — when he saw, partly hidden behind the crown of a pine grove on his left, two identical horizontal red rectangles of light. Slowing down, he watched a third identical rectangle appear to the left of the first two, spaced exactly the same distance apart. The three lights, which looked large to him and sat relatively high in the sky, tracked along the road at his side as he approached a junction ahead — then stopped dead, facing him, just as he brought his own car to a halt a short distance short of the junction to watch them. Closer now, he judged the three lights together spanned roughly two-thirds of his windshield, offset slightly to his left; he had to lean forward over the wheel to take in the whole formation. He described them as looking like truck tail-lights: vivid red, sharply outlined, and each one apparently filled with small moving spheres of light drifting inside its border. Despite gusts strong enough to rock his car on its suspension, the three shapes never wavered. After holding position for several seconds, they moved off to the south, still facing him, still holding their exact formation and appearance, shrinking visually as they went — first at a moderate pace, dropping in apparent height in a series of "bounces" toward the horizon, before suddenly accelerating and vanishing almost instantly upward into the sky, never once passing behind or interacting with the clouds. GEIPAN rated the account's internal consistency "good" and its perceived strangeness high, then took the unusual step of directly contacting the nearby Lanvéoc-Poulmic naval air base and consulting a Marine Nationale drone instructor before testing eight separate conventional hypotheses — and finding that not one of them held up against what the witness described.
Pictured: the sea cliffs (falaises) of the Crozon peninsula, Finistère, France — the rugged Atlantic headland on which the sighting's rural road is located. This is not a photograph of the phenomenon itself; no photograph or video of the three red rectangular lights is known to exist. GEIPAN did produce a video-simulation reconstruction of the witness's account, but it recreates the reported geometry rather than depicting a real recorded image. Photo by Wikimedia contributor Ballaven Patricia, taken March 13, 2016, released under CC BY-SA 4.0 via Wikimedia Commons.
Theoretical Alignment
This case aligns most strongly with the Extraterrestrial Hypothesis among the frameworks this site tracks, on the strength of a structured, geometrically rigid craft-like object — three fixed, evenly spaced, identically shaped light-forms holding a stationary formation against gale-force wind for roughly two minutes, then departing under what the witness described as controlled, patterned motion ending in a burst of near-instantaneous acceleration. It carries modest Interdimensional alignment on the strength of the object's reported "ricochet" descent-then-vanish motion profile — a stepped, bouncing loss of apparent altitude followed by a sudden vertical disappearance is the kind of non-ballistic, non-aerodynamic motion IDH proponents cite as inconsistent with ordinary solid, terrestrial flight. Psycho-Social framing carries comparatively little weight here specifically because GEIPAN's investigators directly tested and eliminated the two most obvious psycho-social candidates — a windshield reflection of an interior cabin light, and a laser or disco-style projector — on the specific grounds that neither could produce three separately positioned, well-defined, internally structured rectangles that remained rigidly fixed across a full two minutes of observation from a moving, then stationary, vehicle.
Background
Understanding why GEIPAN treated a single-witness account with this level of institutional follow-through — direct outreach to a naval air base, a dedicated consultation with a military drone specialist, and formal testing of eight distinct hypotheses — requires understanding both the specific military geography of the Crozon peninsula and the investigative apparatus France has built for exactly this kind of report.
Crozon: A Fortified Peninsula at the Edge of Finistère
The commune of Crozon, home to roughly 7,500 residents, sits at the base of the peninsula that bears its name on the western coast of Finistère, in Brittany — a rugged headland of sea cliffs, heathland, and pine groves reaching out into the Atlantic between the Rade de Brest to the north and the Baie de Douarnenez to the south. It is the seat of the Châteaulin arrondissement, birthplace of the actor Louis Jouvet in 1887, and host each August to the Festival du Bout du Monde. Its more consequential distinction, for the purposes of this case, is military: Crozon's own commune territory includes Île Longue, the fortified base of France's strategic nuclear submarine fleet (the Force Océanique Stratégique), and the peninsula as a whole has hosted a dense concentration of French Navy infrastructure since the early twentieth century, including the naval air base at Lanvéoc-Poulmic, roughly five kilometers from Crozon's town center. Few civilian sighting locations on this site sit this close to this much active, dense, and sensitive French military infrastructure at once — a genuinely relevant fact for weighing the drone and helicopter hypotheses GEIPAN's investigators tested directly against the base itself, discussed at length below.
Lanvéoc-Poulmic began life in 1920 as one of thirty-seven seaplane stations built along the French coast for maritime surveillance, was renamed a "Centre d'Aéronautique Maritime" in 1935, and took its current name in 1937. It endured eighteen separate Allied bombing raids between 1940 and 1944 during the German occupation, and after the war became home to the relocated École Navale (French Naval Academy), whose permanent campus was inaugurated by General de Gaulle in 1965. Today the base hosts three embarked helicopter squadrons — Flottille 32F, 33F, and 34F, flying the H160, Caïman Marine, and Dauphin N3 respectively as of the years around this sighting — alongside training and survival schools, operating as a round-the-clock link in the Marine Nationale's Atlantic-facade military and maritime-law-enforcement mission. Any civilian sighting of unusual lights on this specific peninsula inevitably raises the question of whether a nearby naval asset, rather than anything anomalous, is responsible — which is precisely the question GEIPAN's investigators put directly to the base itself, rather than simply assuming an answer.
The peninsula's own deeper history reinforces just how long this stretch of coastline has mattered strategically to whoever has controlled it. Crozon's own territory carries traces of human settlement stretching back to megalithic monument-builders, followed by Gallo-Roman occupation exploiting the natural anchorage of the Rade de Brest, and centuries of Breton maritime life oriented around fishing and coastal trade before the French state's own military footprint began expanding here in earnest in the early twentieth century. During the Second World War, German occupying forces recognized the same strategic value the French Navy had already identified, fortifying the peninsula as part of the wider Atlantic coastal defense network guarding the approaches to Brest; the eighteen Allied bombing raids Lanvéoc-Poulmic endured between 1940 and 1944 were a direct consequence of that fortification. Postwar, rather than winding down, the peninsula's military role deepened further with the Cold War-era decision to base France's nascent strategic submarine deterrent at Île Longue, cementing Crozon's status as one of the most militarily consequential stretches of coastline in metropolitan France relative to its small civilian population. That is the specific landscape — equal parts working fishing communities, exposed Atlantic cliffs, and dense, decades-deep military infrastructure — against which this sighting's driver was navigating a rural road in a January storm.
GEIPAN's Investigative Machinery: A Small Core, a Wide Network
GEIPAN, based at the Centre National d'Études Spatiales (CNES) campus in Toulouse and traceable back to the GEPAN unit founded in 1977 under astrophysicist Claude Poher, and later reorganized as SEPRA under Jean-Jacques Velasco before adopting its current name and structure, operates with a deliberately small permanent staff — two core positions plus limited administration — multiplied through a network of roughly twenty trained volunteer field investigators and around fifteen outside scientific experts spanning astronomy, meteorology, aeronautics, and psychology, who together form the "comité des experts" convened on the agency's most complex cases. GEIPAN's own published statistics put its annual intake at around 500 witness contacts, of which roughly 200 proceed to a fully investigated, published case file; of those, GEIPAN's own figures indicate about 59% end up explained as straightforward misidentifications, with only around 3% remaining unexplained (Type D) across the agency's full history and just 2% over the past decade specifically, a share that has fallen over time as investigative technique, radar-data access, and public astronomical literacy have all improved.
What distinguishes this specific case within that broader picture is not a field reconstruction or a cognitive interview — the file, unlike some other single-witness cases on this site, does not describe either — but the sheer number and specificity of institutional contacts GEIPAN's investigators made in pursuit of a conventional explanation. Rather than treating the nearby naval base as background context to be assumed away, investigators contacted Lanvéoc-Poulmic directly and obtained a specific, dated negative confirmation; rather than treating "drone" as a single undifferentiated hypothesis, they sought out a Marine Nationale drone instructor specifically qualified to assess whether the reported lighting matched any known military drone's signage. That combination — institutional outreach this specific, on a single-witness case with no photograph, video, or corroborating second observer — sits toward the more thorough end of the investigative effort this site has documented across its published GEIPAN files.
Complete Timeline
| Date | Event |
|---|---|
| 1920 | The seaplane station that becomes Lanvéoc-Poulmic naval air base is established on the Crozon peninsula. |
| 1977 | GEIPAN's predecessor, GEPAN, is founded within CNES in Toulouse under Claude Poher. |
| January 13, 2020, early morning (Phase 1) | Sighting begins. Driving a rural road on the Crozon peninsula in storm conditions, the witness sees two identical horizontal red rectangles, partly obscured by the crown of a roadside pine grove on his left. |
| January 13, 2020 (Phase 2) | Slowing his vehicle, the witness sees a third identical red rectangle appear to the left of the first two, spaced at exactly the same interval. |
| January 13, 2020 (Phase 3) | The three-light formation, appearing large and sitting relatively high in the sky, tracks along the road on the witness's left as he continues driving toward a road junction ahead. |
| January 13, 2020 (Phase 4) | The three lights come to a stop directly over the junction, facing the witness; he stops his own car a short distance before the junction to observe them, now judging them to span roughly two-thirds of his windshield. |
| January 13, 2020 (Phase 5, several seconds) | Close observation: the witness leans forward to see the full formation, describing three sharply outlined, vivid-red, truck-taillight-like shapes, each containing small moving spheres of light, that remain perfectly still despite gusts strong enough to rock his car. |
| January 13, 2020 (Phase 6) | The three lights depart southward, still facing the witness and holding their exact formation and appearance, shrinking in apparent size — first descending in apparent height via a series of "ricochet" bounces toward the horizon, then suddenly accelerating and vanishing almost instantly upward into the sky, never passing behind or interacting with the clouds. Total observation duration: approximately 2 minutes. |
| Following the sighting | The witness reports the sighting to GEIPAN, providing an account investigators would later rate 0.65 for consistency ("good") and note as unusually coherent and detailed. |
| During the investigation | GEIPAN's expert committee (comité des experts) reviews the case in depth, formulating eight distinct hypotheses: military or civilian drone(s), a helicopter, a lighthouse or semaphore beam, a wind-turbine warning beacon, a laser or discotheque projector, a windshield reflection of an interior cabin light, a conventional aircraft, and the reflective plates/position lights of a truck. |
| During the investigation | Investigators deepen the three most plausible-seeming candidates — drones, helicopter, lighthouse/semaphore — while eliminating and documenting the remaining five on independent grounds. |
| During the investigation | GEIPAN contacts the Lanvéoc-Poulmic naval air base directly; the base confirms none of its own aircraft were airborne at the time, and that no military drones were flying that night. |
| During the investigation | A Marine Nationale drone instructor is consulted and states that the reported lighting configuration matches no signage scheme used by any known military drone. |
| During the investigation | A visual reconstruction of the witness's account is produced using a video-simulation tool, intended for publication alongside the full investigation report. |
| Following the full investigation | Final classification. GEIPAN concludes that none of the eight hypotheses tested "gives satisfaction" against the testimony, and classifies the case « D », an unexplained phenomenon, catalogued as case ID 2020-01-50998, with a strangeness rating of 0.55 and a consistency rating of 0.65. |
Detailed Narrative
Two Rectangles, Then Three: The Opening Phase
The witness's account begins in motion, on a small country road crossing the Crozon peninsula, in weather GEIPAN's own file describes plainly as bad: strong wind gusts, a low cloud ceiling, and rain. Driving early in the morning — the precise minute was never published in GEIPAN's own case summary — he noticed, partly masked by the crown of a pine grove standing beside the road on his left, two identical horizontal rectangles of red light. The detail of partial obstruction by real, describable foreground vegetation is itself a small but genuine piece of corroborating context: a witness inventing or badly misremembering a sighting is less likely to spontaneously include a specific, checkable visual obstruction that complicates rather than clarifies what he saw. As he slowed his vehicle to get a better look, a third identical red rectangle appeared to the left of the first two, spaced at what he judged to be exactly the same interval as the gap between the original pair — suggesting, if accurately perceived, a coherent three-unit structure or formation rather than three independently drifting lights that happened to align by chance.
The three-light formation appeared large to the witness and sat relatively high in the sky. Rather than remaining static, it began tracking along the road, staying on his left side as he continued driving toward a junction of roads a short distance ahead. This tracking behavior — the lights maintaining a roughly constant angular position relative to the moving vehicle, rather than the vehicle simply passing beneath a fixed object, or the object crossing the sky independent of the car's motion — is a detail GEIPAN's investigators would need to account for however they ultimately explained the incident, since a genuinely fixed object (a lighthouse beam, a wind-turbine beacon) cannot itself "track" a moving vehicle in the way the witness described, while a genuinely mobile object doing so raises its own separate strangeness.
The Stop at the Junction: A Sustained, Close-Range Observation
The case's most evidentially significant phase is also its longest and closest. As the witness approached the junction, the three lights came to a halt directly over it, facing him; he, in turn, brought his own car to a stop a short distance before reaching the junction specifically in order to observe them at length. This is the point in the account where casual passing perception gives way to sustained, close, deliberate observation — the witness was no longer glancing at something in his peripheral vision while driving, but stationary, facing the phenomenon directly, for what the case file describes as several continuous seconds within a total observation window of roughly two minutes.
From this stopped, close vantage point, the witness offered his most specific measurements and descriptions. He judged the three-light formation, taken together, to span roughly two-thirds of the width of his windshield, positioned slightly to his left rather than dead ahead — a detail suggesting the object's exact bearing rather than a vague "in front of me" recollection. He had to lean forward in his seat, over the steering wheel, to take in the full formation at once, implying an apparent angular size large enough that it did not comfortably fit within a single relaxed sightline. His description of the individual lights themselves was specific and unusual: each resembled the tail-light cluster of a large truck, vivid red, sharply and cleanly outlined rather than diffuse or glowing, and — the account's single strangest visual detail — each rectangle appeared to contain small spheres or points of light that moved within its border, as though the illumination inside each panel were not uniform or static but composed of discrete, mobile internal elements. Despite wind gusts strong enough to physically rock his stationary car on its suspension, all three shapes held perfectly rigid, unmoving formation throughout this entire close-range phase.
Departure: A Bouncing Descent, Then Near-Instant Disappearance
After holding position for several seconds under close observation, the three lights began to move again, this time departing to the south while still facing the witness directly, without ever breaking their exact mutual spacing or individual appearance. Rather than a single continuous motion, the witness described the departure as unfolding in two distinct tempos. First, the formation moved at what he judged a moderate pace, visually shrinking in apparent size while dropping in apparent height through a series of discrete steps or "bounces" — GEIPAN's own file records the witness's own term, "ricochets" — progressively lowering toward the horizon rather than climbing or moving in a smooth, continuous arc. Then, without further warning, the formation accelerated sharply and vanished almost instantaneously upward into the sky. GEIPAN's file specifically notes that this final disappearance occurred without the lights ever passing behind, into, or otherwise visibly interacting with the low cloud ceiling present that morning — a detail investigators would need to weigh carefully, since a genuine departure into or through actual cloud cover is one of the more mundane ways a light source can simply become obscured from view, and its explicit absence here removes that easy explanation from consideration.
A Uniquely Institutional Investigation
What sets this case apart from most single-witness GEIPAN files this site has published is not the richness of the witness's own account, though that is considerable, but the specific character of the investigative response it triggered. GEIPAN's expert committee formulated eight distinct hypotheses in total: one or more drones, a helicopter, a lighthouse or maritime semaphore beam, a wind-turbine obstruction-warning beacon, a laser or discotheque-style light projector, a reflection on the windshield of a light source inside the witness's own vehicle, a conventional fixed-wing aircraft, and the reflective plates or position lights of a truck. Of these eight, investigators judged three — drones, helicopter, and lighthouse/semaphore — substantial enough to warrant genuine deepened investigation, while the remaining five were examined and eliminated on more straightforward grounds, discussed individually below in the Conventional Explanation Candidates section.
The deepened drone hypothesis led investigators to Lanvéoc-Poulmic itself, the nearby naval air base whose proximity to the sighting site makes it the single most obvious candidate source for any unusual aerial activity on the peninsula that morning. The base's own confirmation — that none of its aircraft were airborne at the relevant time, and that no military drones were flying that night either — is a specific, dated, institutionally sourced negative result, categorically different from a general assumption that "the base probably wasn't involved." GEIPAN's investigators went further still, consulting a Marine Nationale drone instructor specifically qualified to assess lighting configurations used by military drone platforms; that instructor's professional judgment was that the reported appearance — three fixed, sharply outlined red rectangles containing internally mobile points of light — corresponds to no lighting signage scheme used by any drone type known to that instructor. This is a genuinely rare category of evidence on this site: not merely an absence of a matching radar contact, but a positive, expert, named-role professional opinion that the reported visual signature itself doesn't match the relevant technology category at all.
What the Case File Does Not Contain
It is worth being explicit about what GEIPAN's published summary does not include, since the gaps are themselves informative. The file does not name the exact road or crossroads involved, does not give a precise time of day beyond "early morning," and does not disclose the witness's occupation, all consistent with the agency's standard anonymization and location-generalization practice rather than any indication of a thinner or less-documented file. The file also does not describe a formal cognitive interview of the kind applied in some other single-witness cases on this site, nor a second field visit beyond the one reconstruction roughly a year after the sighting. Readers comparing this case against others on this site should weigh its investigative profile on its own specific terms — heavy on direct institutional outreach, comparatively light on the repeated witness re-interviewing seen elsewhere — rather than assuming every GEIPAN file follows an identical procedural template.
A Note of Institutional Caution
GEIPAN's own file, notably, does not treat this thoroughness as license for full confidence in its own D classification. The technical questionnaire that produced the account's rich detail was completed only five months after the sighting itself, and the field reconstruction of the site's geometry was not carried out until roughly a year after that — delays GEIPAN's investigators explicitly flag, per their own accumulated institutional experience, as capable of disturbing elements of a witness's testimony over time. This is a candid, self-critical note rather than a reflexive endorsement, and it matters for how the case should be weighed: GEIPAN's own investigators are on record acknowledging that the very account they ultimately rated "good" for consistency was reconstructed and formally documented well after the kind of short window in which memory distortion risk is lowest, even as they still found no explanation that accounted for what was reported.
Key Witnesses
The Crozon Peninsula Primary Witness (identity withheld per GEIPAN's standard anonymization practice)
A driver on a rural road on the Crozon peninsula who observed the full sequence — two lights becoming three, a sustained close-range stop at a road junction, and a two-phase departure ending in near-instantaneous disappearance — over roughly two minutes during a January storm. He later provided GEIPAN with a detailed technical questionnaire, completed five months after the sighting, from which investigators derived a consistency rating of 0.65 ("good") and a strangeness rating of 0.55. A field reconstruction of the sighting's geometry followed roughly a year after that. GEIPAN's own file explicitly flags the length of both delays as a factor that can, per the agency's general experience, disturb elements of any witness's testimony, even while ultimately finding no conventional explanation that reduces the account's underlying strangeness.
The Five Observables Assessment
The Five Observables framework, developed by the Pentagon's Advanced Aerospace Threat Identification Program to describe performance characteristics beyond known conventional technology, is applied here for comparative consistency across the site's case files. This case registers one clean positive match, two genuinely ambiguous speculative ratings, and one notable evidentiary gap distinct from comparable cases on this site.
Assessment: This case's cleanest positive observable is Instant Acceleration: the witness's own account distinguishes a first, slower phase of "ricochet"-style descent from a distinctly separate, sudden final burst of upward acceleration and disappearance, mapping directly onto AATIP's canonical category with little of the interpretive hedging that qualifies this case's other observables. Anti-Gravity Lift is rated speculative rather than present specifically because the witness never described any visible lift mechanism — only the striking fact that gale-force wind, strong enough to physically move his own vehicle, produced no visible effect whatsoever on three ostensibly airborne light-forms holding a fixed geometric arrangement. Low Observability stands out for a different reason than in comparable radar-checked cases elsewhere on this site: rather than a clean negative result (Montluel) or a partial, mistimed one (Sainte-Foy-de-Peyrolières), this case has no radar data at all, since GEIPAN's investigation proceeded via direct institutional confirmation (the naval base, the drone instructor) rather than a radar-track reconstruction — a genuinely different, and in some ways less complete, evidentiary gap.
Evidence Assessment
Supporting an Anomalous Origin
- An unusually detailed, internally consistent single-witness account earning a "good" (0.65) GEIPAN consistency rating and a high perceived-strangeness rating (0.55)
- GEIPAN formulated and tested eight distinct hypotheses, deepening the three most plausible, without settling on any explanation that survived scrutiny
- A direct institutional check with the nearby Lanvéoc-Poulmic naval air base returned a specific, dated negative confirmation: no aircraft airborne, no drones flying that night
- A Marine Nationale drone instructor's expert professional opinion found the reported lighting signature matched no known military drone platform
- The lighthouse/semaphore hypothesis was independently judged very unlikely on three separate technical grounds (insufficient beam power, absence of an expected visible beam under the rectangles in rainy conditions, and the witness's own prior familiarity with the area's actual lighthouse lighting)
- The reported final acceleration and near-instantaneous disappearance maps cleanly onto the AATIP Five Observables framework's "instant acceleration" category
Supporting a Conventional/Mundane or Inconclusive Origin
- This remains a single, uncorroborated witness account with no second observer, no photograph, and no video of the phenomenon itself
- The technical questionnaire was completed five months after the sighting, and the field geometry reconstruction roughly a year after that — delays GEIPAN's own investigators explicitly flag as a genuine risk to testimony accuracy
- No radar check of any kind was obtained, unlike several comparable cases on this site, leaving a real, unaddressed gap in the instrumented record rather than a confirmed absence of conventional air traffic
- The base and drone-instructor consultations, while genuinely negative, rest on institutional self-reporting rather than independent physical verification
Why "Open"
This case is stamped Open rather than Credible, Disputed, or Debunked because GEIPAN's own investigation reached precisely that state: an unusually institution-facing, multi-hypothesis inquiry that directly contacted the single most obvious potential conventional source — a nearby naval air base — and came back with a specific negative answer, without ever securing the kind of second witness, photograph, or instrumented data that would justify a stronger designation in either direction. The case earns real, substantial credit for the specificity of its negative institutional findings: a dated confirmation from Lanvéoc-Poulmic that no aircraft or drones of its own were active, a named-role military drone instructor's professional judgment that the lighting signature matches no known platform, and a lighthouse/semaphore hypothesis eliminated on three independent technical grounds rather than dismissed casually. Eight formally tested hypotheses on a single-witness case, with three genuinely deepened rather than merely listed, sits toward the more thorough end of the investigative effort this site has documented.
Set against that, "Credible" would overstate what the public record actually supports: this remains, in the end, a single witness with no photographic or video evidence and no second observer, and unlike several comparable cases on this site, no radar reconstruction was ever obtained at all — a real evidentiary gap, not merely an incomplete one. GEIPAN's own investigators additionally flagged, honestly and without prompting, that the five-month and roughly one-year delays before the questionnaire and field reconstruction respectively are the kind of gap that can disturb testimony, tempering how much weight even a "good" consistency rating should carry. A case whose own investigating agency reached out directly to a military installation, obtained a specific negative answer, tested eight hypotheses in total, and still concluded that nothing it could verify explains what was reported — while candidly flagging its own investigation's timing limitations — is exactly the state "Open" is meant to capture: genuinely, thoroughly investigated, and genuinely unresolved.
Sensor & Instrumentation Detection Profile
This case's instrumented evidence is entirely institutional rather than radar- or sensor-based. No air-traffic or air-defense radar reconstruction of any kind was sought or obtained for this sighting, a genuine point of difference from several comparable GEIPAN cases on this site (Montluel's negative CAPCODA/CNOA check, Fauville-Évreux's positive uncorroborated contact, and Sainte-Foy-de-Peyrolières's partial, mistimed Toulouse-Blagnac data), each of which at least attempted some radar-based cross-check. In place of radar, GEIPAN's investigators substituted direct institutional confirmation: a formal query to the Lanvéoc-Poulmic naval air base, which confirmed no aircraft of its own were airborne and no military drones were flying that night, and a consultation with a Marine Nationale drone instructor, who assessed the reported lighting signature against known military drone platforms and found no match. No photograph, video, or audio recording of the phenomenon itself exists; the only visual record connected to the case is GEIPAN's own after-the-fact video-simulation reconstruction of the witness's verbal account, intended for publication alongside the full investigation report rather than as documentary evidence of the sighting itself.
Environmental & Geospatial Context
The sighting occurred during genuinely poor weather — strong wind gusts, a low cloud ceiling, and rain — conditions GEIPAN's investigators treated as directly relevant to at least two of the hypotheses tested. The absence of any visible beam beneath the three light-forms, despite rainy conditions in which a genuine lighthouse or powerful ground-based beam would be expected to scatter visibly off falling rain, weighed directly against the lighthouse/semaphore hypothesis. Separately, the wind's strength — sufficient to rock the witness's own stationary vehicle — became a key data point cutting against a helicopter explanation, since a helicopter large and close enough to produce the reported angular size should, under GEIPAN's investigators' own reasoning, have been more rather than less audible in gusty conditions carrying sound toward the witness, not less.
Geographically, the sighting occurred on the Crozon peninsula, a stretch of Finistère coastline carrying an unusually dense concentration of French military infrastructure for a rural area of its size: the naval air base at Lanvéoc-Poulmic roughly five kilometers away, and the highly sensitive Île Longue strategic submarine base within Crozon's own commune territory. This military density cuts in two directions at once for evaluating the case: it means genuine, mundane military air activity (a helicopter transit, a drone exercise) is a background possibility any investigator must take seriously here more than in a purely civilian area — which is exactly why GEIPAN contacted the base directly rather than assuming an answer — but it also means that base's specific, dated negative confirmation carries real evidentiary weight precisely because a false or careless answer would have been unusually easy to check against, given how much of the surrounding area's air traffic that single base could plausibly account for.
The peninsula's exposure to Atlantic storm systems is itself an ordinary, year-round feature of the local climate rather than an unusual event in its own right — Finistère's western coastline routinely absorbs the first landfall of winter depressions tracking in off the North Atlantic, and January is climatologically among the windiest and wettest months in the department. That baseline matters for weighing the case in two opposite directions at once. On one hand, it means the specific weather conditions present that morning were unremarkable for the season and region, and cannot themselves be treated as an unusual contributing factor requiring separate explanation. On the other, it means the witness — as a resident of an area accustomed to this kind of weather, and, per GEIPAN's own reasoning on the lighthouse hypothesis, already familiar with how the area's genuine navigational lighting looks under comparable storm conditions — was not an inexperienced observer encountering an unfamiliar visual environment for the first time. A witness who already knows what the local coastline's real lights, beacons, and beams look like in bad weather, and who still could not identify what he was seeing as any of them, is a materially different kind of report than one from a visitor unfamiliar with the area's ordinary nighttime and storm-lit skyline.
Observer Credibility & Occupational Profile
The witness's occupation is not disclosed in GEIPAN's published case file, consistent with the agency's standard anonymization practice, but the file's own procedural record documents a credibility profile built from behavior: a willingness to complete GEIPAN's full technical questionnaire in specific, phase-differentiated detail (partial obstruction by a real pine grove, an exact estimate of angular width relative to his windshield, a specific description of internal structure within each light), and evident cooperation with whatever field reconstruction process followed roughly a year later. GEIPAN's investigators themselves rated the resulting account's consistency 0.65, "good" — a rating earned despite, not because of, the case's evidentiary limitations, since a single uncorroborated account with no photograph must earn that rating purely on internal coherence and specificity rather than any external cross-check.
The case's structural limitation is the one shared by the majority of single-witness UAP reports generally: however internally consistent and specific one account is judged to be, it cannot by itself carry the same evidentiary weight as independent corroboration from a second observer, a photograph, or an instrumented data source. That limitation is compounded here by GEIPAN's own candid acknowledgment that the five-month gap before the technical questionnaire, and the roughly year-long gap before the field reconstruction, are exactly the kind of delays that can, per the agency's own accumulated experience, disturb elements of a witness's testimony — a genuine, self-identified caveat this case carries that several faster-documented cases on this site do not.
Historical Precedents & Archive Matches
Comparing this case against the other GEPAN/GEIPAN-investigated French sightings already published on this site highlights both its distinctive investigative profile and where its evidentiary strength sits relative to comparable files.
Sainte-Foy-de-Peyrolières (2018) — Two Different Models of Investigative Thoroughness
The Sainte-Foy-de-Peyrolières Five-Phase Sighting (Open, 47/100) offers the closest comparison in overall investigative intensity: another single-witness, no-photograph case that received unusually thorough treatment, there via two separate expert-committee reviews and a formal field reconstruction with cognitive interview. This case's own thoroughness took a different, more institution-facing form — rather than re-interviewing the witness twice, GEIPAN's investigators here reached outward, contacting a naval air base directly and consulting a named military drone specialist. Both approaches reflect real investigative effort; neither produced a conventional explanation that survived scrutiny.
Montluel (2022) and Fauville-Évreux (1978) — The Radar Check This Case Never Had
The Montluel cloud-distorting object (Open, 46/100) and the Fauville-Évreux radar sighting (Open, 48/100) both involved a genuine air-defense or air-traffic radar cross-check — a clean negative result in Montluel's case, a positive uncorroborated contact in Fauville-Évreux's. This case's investigation proceeded along an entirely different evidentiary track: no radar reconstruction of any kind was sought, replaced instead by direct confirmation from the nearby naval base and its own drone specialist. Whether an institutional confirmation of this kind is weighed as comparably strong to a negative radar result, or as a genuinely different and less independently verifiable category of evidence, is a fair question this case's own file leaves open, and this site does not attempt to resolve it definitively in either direction.
Montayral (2015) and Bornel (2014) — Single-Witness Cases at Different Consistency Levels
The Montayral aerodrome object (Open, 28/100) is the site's other single-witness, no-photograph case with the least additional corroboration, its consistency rated only "just sufficient." This case's own 0.65 consistency rating sits meaningfully higher, though still below Sainte-Foy-de-Peyrolières's 0.72 — a genuine middle position reflecting a detailed, specific account that nonetheless lacks that case's multi-phase structural richness. The Bornel conservatory fireball (Open, 51/100), by contrast, benefits from two corroborating witnesses and an inconclusive but real photograph — the kind of multi-source corroboration this Crozon file, like most single-witness cases on this site, simply does not have, however unusually thorough its underlying investigation was.
Material Analysis
This case involves no physical or trace evidence of any kind — no landing marks, no recovered material, no soil or vegetation sample, and no biological or physiological effect on the witness beyond his own visual observation. Nothing in the account places the object at ground level or within the witness's physical reach at any point; the three light-forms remained airborne throughout, at a distance the witness himself judged as relatively high in the sky even during the closest phase of observation. There is accordingly nothing to sample, test, or physically examine after the fact — only the witness's own detailed verbal and questionnaire-based description of what he observed, and the specific negative confirmations GEIPAN obtained from the nearby naval base and its drone instructor.
Supporting Non-Terrestrial Origin
- No physical evidence exists to test in the first place, so no terrestrial material-composition finding contradicts an anomalous interpretation either
- The object's reported rigid formation-holding against gale-force wind, its internally structured "moving spheres" lighting pattern, and its final near-instantaneous acceleration, if accurately perceived, describe a performance envelope and physical construction outside any conventional aircraft, drone, or fixed light source GEIPAN identified
Supporting Terrestrial Origin
- The complete absence of any physical trace is itself consistent with a purely distant, airborne phenomenon that never approached the ground closely enough to leave one
- No mechanical, electromagnetic, or biological secondary effect on the witness, his vehicle's systems, or any nearby structure is recorded anywhere in the file
The Conventional Explanation Candidate: Eight Hypotheses, None Satisfactory
GEIPAN's own file is unusual among the cases published on this site in tying its final classification directly to a specific number of tested hypotheses: eight, spanning aviation, maritime infrastructure, entertainment lighting, optical illusion, and ground-vehicle misidentification. Three — drones, a helicopter, and a lighthouse or maritime semaphore beam — were judged substantial enough to deepen; the remaining five were examined and set aside on more immediate grounds. None, in GEIPAN's own words, "gives satisfaction" against the witness's testimony.
Military or Civilian Drone(s)
The most directly tested hypothesis, given the sighting's proximity to Lanvéoc-Poulmic. The base's own confirmation that no military drones were flying that night, combined with a Marine Nationale drone instructor's professional judgment that the reported lighting — three fixed, sharply outlined red rectangles with internally moving light points — matches no signage scheme used by any known military drone platform, leaves this hypothesis without institutional or technical support. A civilian drone capable of producing this specific appearance, holding rigid formation with two other identical units against gale-force wind, for two full minutes, at the apparent altitude and size the witness described, is not accounted for by any drone technology GEIPAN's investigators identified as available in 2020.
Helicopter
The angular size the witness reported — the formation spanning roughly two-thirds of his windshield at closest approach — implies, at any plausible real-world altitude, a position close enough to the witness that a helicopter of matching size should have been clearly audible, especially given wind conditions that, if anything, should have carried engine noise toward him rather than away. No sound of any kind was reported at any point during the two-minute observation. The hypothesis also does not account for three separate, identically shaped, rigidly fixed red rectangles rather than a single fuselage with conventional position lighting.
Lighthouse or Maritime Semaphore Beam
Judged, after deepened investigation, very unlikely on three independent grounds: no lighthouse in the area is powerful enough to produce three fixed, well-delimited red rectangles sustained for a full two minutes; a genuine lighthouse or semaphore beam of the required power should have produced a visible beam beneath the rectangles under the evening's rainy conditions, which the witness did not report; and the witness, familiar with the area, would plausibly have recognized the area's actual lighthouse lighting under weather conditions he had already experienced before.
Wind-Turbine Warning Beacon
Eliminated on straightforward grounds: wind-turbine obstruction beacons are typically single, unstructured points or flashing lights at a fixed location, not three coordinated, internally structured rectangles capable of tracking a moving vehicle and then departing at speed.
Laser or Discotheque-Style Projector
Eliminated for lacking any plausible fixed or mobile source capable of producing three separately positioned, sharply outlined rectangles — rather than beams, points, or patterns — visible steadily for two minutes from a moving and then stationary vehicle at the reported distance and angular size.
Reflection on the Windshield of an Interior Cabin Light
A genuine windshield reflection is inherently tied to the observer's own head position and the geometry of the vehicle's interior, and would not be expected to remain fixed, structured, and consistent across both a moving-vehicle phase and a later stationary, close-range observation phase, nor to appear to move and hold formation independently as the vehicle itself changed position and orientation approaching the junction.
Conventional Aircraft
A single aircraft, considered as the simplest possible candidate, cannot account for three separately positioned, identically appearing rectangular light-forms holding rigid mutual spacing, nor for the total absence of any engine sound despite favorable wind conditions for sound propagation and an implied close range.
Reflective Plates or Position Lights of a Truck
Eliminated on straightforward geometric grounds: a truck's reflectors or position lights are fixed to a ground-level vehicle and cannot appear "relatively high in the sky," track a separate moving vehicle from that elevation, or depart upward and vanish into the sky at the end of the encounter.
Military or Civilian Drone(s)
ImplausibleThe nearby naval base confirmed no military drones were flying; a Marine Nationale drone instructor found the lighting matches no known military drone platform.
Helicopter
ImplausibleThe implied close range for the reported angular size should have produced audible engine noise, especially in gusty wind; none was reported.
Lighthouse / Maritime Semaphore
Very UnlikelyNo local lighthouse is powerful enough for this effect; no beam was visible under the rectangles despite rain; the witness knew the area's real lighthouse lighting already.
Wind-Turbine Warning Beacon
ImplausibleA single fixed beacon cannot produce three coordinated, structured, vehicle-tracking rectangles.
Laser / Discotheque Projector
ImplausibleNo plausible source produces three separately positioned, sharply outlined rectangles sustained for two full minutes.
Windshield Reflection
ImplausibleA reflection tied to the observer's head position would not hold a fixed, structured appearance across both a moving and a stationary observation phase.
Conventional Aircraft
ImplausibleCannot account for three identical, rigidly spaced rectangles or the total absence of engine sound.
Truck Reflectors / Position Lights
ImplausibleGround-level fixtures cannot appear high in the sky, track a separate vehicle, or depart upward into the sky.
Related Cases
Confirmed Documents & Official Records
The official French government investigation summary: the witness's full account, the eight tested hypotheses, the Lanvéoc-Poulmic and drone-instructor consultations, and the final Class D determination with published strangeness (0.55) and consistency (0.65) ratings.
GEIPAN's full field investigation report, including the technical questionnaire results, the hypothesis-by-hypothesis analysis, and the institutional correspondence with Lanvéoc-Poulmic.
GEIPAN's own published definitions for classifications A, B, C, D, D1, and D2, and the strangeness/consistency evaluation criteria applied to this case.
GEIPAN's own account of its investigative structure — the roughly twenty trained field investigators and fifteen-member scientific expert committee — and its statistical case outcomes, providing institutional context for this case's unusually institution-facing investigation.
Further Reading — Recommended Literature
UFOs: Generals, Pilots, and Government Officials Go on the Record
Kean's survey of credentialed witness testimony and official investigative bodies worldwide provides useful comparative context for weighing GEIPAN's own institution-facing investigative approach against how other national UAP investigation programs operate.
Wonders in the Sky: Unexplained Aerial Objects from Antiquity to Modern Times
A catalog of historical aerial anomaly reports useful for weighing how genuinely unusual this case's specific rigid, formation-holding, structured-light phenomenon is against the broader historical pattern of reported aerial phenomena.
The UFO Evidence
The foundational documentary approach to cataloging and cross-referencing witness testimony that GEIPAN's own case-file methodology, including its formal consistency and strangeness ratings, descends from in spirit.
The UFO Experience: A Scientific Inquiry
Hynek's own multi-hypothesis evaluation framework anticipates the eight-hypothesis, institution-consulting methodology GEIPAN applied to this exact case, making it a natural comparative reference.
The Close Encounters Man: How One Man Made the World Believe in UFOs
A biography of Hynek documenting the emergence of rigorous, elimination-based case investigation methodology, directly reflected in GEIPAN's eight-hypothesis handling of this case.
Essential Viewing
Sources
- GEIPAN (CNES), "CROZON (29) 13.01.2020", Case 2020-01-50998 — official case summary
- GEIPAN (CNES), Compte Rendu d'Enquête, Crozon (29) 13.01.2020 — full investigation report (PDF)
- GEIPAN (CNES), official case archive & public portal — geipan.fr
- GEIPAN (CNES), Glossaire du GEIPAN — official classification definitions, A/B/C/D/D1/D2
- GEIPAN (CNES), "Missions, méthodes et résultats" — investigative structure, expert committee methodology
- Wikipédia (français), "Base d'aéronautique navale de Lanvéoc-Poulmic" — base history, founding, current squadrons
- Wikipedia (English), "Crozon" — commune geography, population, Île Longue reference
- World Airport Codes, "Lanvéoc-Poulmic Air Base" — ICAO LFRL, coordinates, facility data
- Ministère des Armées (Archives), "Base d'aéronautique navale Lanvéoc-Poulmic" — official Marine Nationale historical record
- Mémorial national des marins morts pour la France, "Base aéronavale Lanvéoc-Poulmic" — wartime and institutional history
- This site, case-files/sainte-foy-de-peyrolieres-five-phase-sighting.html — comparative single-witness GEIPAN Class D case with twice-reviewed expert committee
- This site, case-files/montluel-cloud-distorting-object.html — comparative GEIPAN case with a clean negative radar check
- This site, case-files/montayral-aerodrome-object.html — comparative single-witness GEIPAN D1-family case
- This site, case-files/fauville-evreux-radar-sighting.html — comparative GEPAN/GEIPAN positive radar case
- This site, case-files/bornel-conservatory-fireball.html — comparative two-witness GEIPAN case with inconclusive photograph
- Wikimedia Commons, File: Falaises de Crozon.jpg, photo by "Ballaven Patricia" (March 13, 2016), CC BY-SA 4.0 — licensing and attribution
- YouTube (C dans l'air / France Télévisions), "Geipan : la France aussi s'intéresse aux ovnis" — oEmbed-verified video, general GEIPAN institutional coverage
- This site, key-figures.html, entry on Claude Poher — GEPAN founding and early institutional history
- This site, key-figures.html, entry on Jean-Jacques Velasco — SEPRA leadership and methodology evolution
- This site, locations/index.html, entry on Toulouse — CNES/GEPAN/SEPRA/GEIPAN institutional headquarters
See something inaccurate? Have additional information or sources?
Suggest an Edit →Last Updated: August 2026
