1. Home
  2. Knowledge Base
  3. I - Data and Sensors
  4. 60. Archives and Declassification
  5. GEIPAN: France’s Official UAP Unit

GEIPAN: France’s Official UAP Unit

GEIPAN (Groupe d’Études et d’Informations sur les Phénomènes Aérospatiaux Non-identifiés) is France’s official UAP investigation unit, operating under CNES (the national space agency) since 1977. With a mission to collect, analyze, and publicly archive reports of Unidentified Aerospace Phenomena (UAP), GEIPAN is renowned for its scientific rigor and transparency. The agency uses a structured classification system (A/B/C/D1/D2) based on weirdness and consistency, with only 2–3% of cases remaining unexplained (D1/D2) after investigation, many involving credible witnesses or multi-sensor data.

GEIPAN’s methodology includes cognitive interviews, on-site investigations, and the use of IPACO®, a forensic image analysis suite. Supported by France’s Gendarmerie and Air & Space Force, GEIPAN can cross-check radar data and benefits from multi-agency governance via its COPEIPAN steering committee. All validated case files are anonymized and made available online.

The office is also a global model for openness, publishing over 3,200 classified cases and engaging with researchers through conferences like CAIPAN. While most sightings are attributed to natural or man-made phenomena, GEIPAN continues to highlight a small, persistent set of anomalous cases, emphasizing the importance of continuous, methodical investigation.

What is GEIPAN?

GEIPAN (Groupe d’Études et d’Informations sur les Phénomènes Aérospatiaux Non-identifiés) is the French government’s office for Unidentified Aerospace Phenomena (UAP). It sits inside CNES, the French space agency, and has three core missions: collect, analyze, and archive UAP reports, and inform the public. CNES created the program in 1977 (as GEPAN), restructured it in 1988 (as SEPRA), and relaunched it in 2005 as GEIPAN with a public transparency mandate, including a searchable database of cases and downloadable investigation files. CNES

GEIPAN deliberately uses the term UAP (in French, PAN) instead of “UFO,” noting that many sightings are not of solid “objects” and that “UFO” carries unnecessary cultural baggage. Geipan

Since January 15, 2024, GEIPAN has been led by Frédéric Courtade. Oversight is provided by COPEIPAN, a steering committee with representatives from civil and military authorities (Gendarmerie, Police, Air & Space Force) and scientific bodies (CNES, CNRS, Météo-France, psychologists, etc.). Geipan

GEIPAN by the Numbers (as of August 2025)

Live statistics (dynamic):

  • 3,240 total classified cases published on the site (as of August 21, 2025).
  • Classification distribution: A 27.19%, B 38.73%, C 30.93%, D 2.16%. Sub-breakdown shows D1 0.99% and D2 displayed as 0.00% (likely a rounding/display artifact, see notes below). Geipan

Long-term activity (cumulative program totals):

  • Across the last 40 years, 9,724 testimonies (~5,300 cases) analyzed; ~10% led to on-site investigations. GEIPAN estimates ~3–3.4% of sightings remain unexplained after investigation, with a drop toward ~2% over the last decade due to re-appraisals and better tools. In 2017 alone, ~50 legacy D cases were re-investigated and explained. Geipan

How to interpret the “D / D1 / D2” split: the live statistics are “dynamic,” reflecting only published, classified cases at a given moment; D1 and D2 are subsets of D. GEIPAN retains the D1/D2 distinction in methodology (see below), even if the widget sometimes rounds the D2 share to 0.00%. Geipan

Governance, Partners, and What “Official” Really Means

GEIPAN operates inside CNES with a small core team and a network of vetted volunteer field investigators known as IPN (Intervenants de Premier Niveau), who support on-site work under defined rules. Recruitment has been paused/resumed at times to maintain standards. Geipan

Two key state partners are embedded in the governance loop:

  • Gendarmerie Nationale: a major intake channel for reports because of its territorial presence; it can take formal statements (procès-verbaux) and forward them to GEIPAN. gendarmerie.interieur.gouv.fr
  • Armée de l’Air et de l’Espace (Air & Space Force): GEIPAN can request radar checks for specific times/locations. Both the Air Force and Gendarmerie sit on the steering committee. Geipan

The office stresses public service and transparency: archives have been online since 2007, and the database continues to expand with anonymized reports, investigation files, and conclusions.

From Sighting to Case File: GEIPAN’s Data Pipeline

1) Witness intake.
Members of the public can report directly via GEIPAN’s site using structured questionnaires, or file a report with the Gendarmerie/Police (who then forward it). The witness forms are detailed (location, azimuth/elevation if known, duration, environmental conditions, sketches, photos/videos) and include privacy & retention notices. Notably, GEIPAN states that witness data associated with testimonies are retained for up to 99 years, consistent with archival practice and GDPR notices in the form. Geipan

2) Pre-analysis & triage.
Staff and experts screen for obvious identifications (astronomy, satellites, atmospheric optics, aircraft, drones, re-entries, etc.) using checklists and external resources (their FAQ even links to Heavens-Above and other tools witnesses can try themselves). geipan.fr

3) Investigation.
Where warranted- especially for D1/D2 candidates, GEIPAN deploys on-site work: cognitive interviews with witnesses, scene reconstructions, environmental checks, and, if available, multi-sensor corroboration (radar queries to the Air Force, cross-reports). The office explicitly states that D cases are periodically re-evaluated when new data appear. Geipan

4) Imagery & video workflow.
For photo/video submissions, GEIPAN uses IPACO®, an image authentication and analysis suite co-financed by CNES-GEIPAN since the 2010s. IPACO checks for signs of manipulation (via metadata/EXIF and forensic tests), optical artifacts, and then assists with geometric/photometric analysis to support identifications or strengthen D-class cases. Geipan

5) Publication & anonymization.
GEIPAN publishes anonymized reports and conclusions in its database, along with investigation dossiers (often PDFs). Users can search by region, year, classification, and phenomenon type. Geipan

The Classification System (A/B/C/D1/D2), in Plain Language

Since 2008, GEIPAN has formalized a two-axis evaluation: weirdness (E) vs consistency, yielding five practical outcomes: A, B, C, D1, D2. In short:

  • A – Perfectly identified (e.g., Venus, Starlink train, aircraft on approach).
  • B – Probably identified (a best-fit explanation surpasses alternatives; some uncertainty remains).
  • C – Unidentified / not workable (insufficient data to test hypotheses).
  • D – Unidentified after investigation (robust case despite effort).
    • D1: “strange,” medium consistency (e.g., single credible witness with no sensors).
    • D2: very strange, strong consistency (e.g., multiple independent witnesses and/or imagery/trace evidence).

GEIPAN’s formalism defines weirdness as 1 minus the probability of the strongest prosaic hypothesis. If E > 0.5 and consistency is high, the case may be D1/D2; if E < 0.5, it’s A/B; otherwise C. D cases are re-evaluated periodically. Geipan

What the Data Say (and Don’t Say)

The live stats (Aug 21, 2025) show A+B ≈ 65.9%, C ≈ 30.9%, D ≈ 2.16% (with D1 ≈ 0.99%; D2 displayed as 0.00%). GEIPAN’s broader historical notes report ~3–3.4% D across decades, dropping toward ~2% in recent years as old D cases are re-examined with modern tools. Two important nuances:

  1. Dynamic denominator: the live widget counts published and classified case files; it excludes in-progress dossiers and re-classifications pending publication. That creates epochal effects, recent years initially show fewer D’s because complex cases take longer to resolve and publish. Geipan
  2. D1 vs D2 display: the site treats D1/D2 as sub-bins of D. The current 0.00% display for D2 is likely a rounding or roll-up issue in the widget; GEIPAN still uses and defines D2 formally, and publishes D2 dossiers when supported by multi-witness/sensor evidence. Geipan

A second long-term signal is just as interesting: on-site investigations occur in roughly ~10% of cases, reserved for more complex or promising reports; hoaxes are rare (<1%) according to GEIPAN. Geipan

A Canonical Case: Trans-en-Provence (1981)

Although predating GEIPAN’s current structure, the Trans-en-Provence investigation (GEPAN era) is frequently cited because it involved ground traces, lab analyses (soil and plants), gendarmerie procedures, and extensive documentation, exactly the forensic model GEIPAN still pursues when possible. GEIPAN’s case page summarizes the witness account, physical marks, and sampling by the Gendarmerie within a day of the event. Geipan

Interpretations disagree. The original technical notes reported soil compaction/heating and plant stress gradients; skeptical critiques have proposed mundane alternatives (e.g., tire skids). This makes Trans-en-Provence a useful template for how D-class cases can remain Disputed even when thoroughly documented. 

Events, Outreach, and Research

GEIPAN runs/hosts scientific colloquia, most recently CAIPAN II (Toulouse, Oct 13–14, 2022) to cross-pollinate methods across aviation safety, atmospheric physics, psychology, imaging, and data science. The office regularly publishes news posts about explanatory patterns (e.g., pilot reports involving cockpit “flashes”) and methodology tools (IPACO®). france-science.com

How GEIPAN Compares Internationally

  • Transparency first. Since 2007, France has published case files routinely, rather than only aggregated statistics or internal summaries. That practice remains rare worldwide, though Canada has begun programmatic updates, and the U.S. AARO provides a federal reporting pathway and high-level reports (with less case-level disclosure). Smithsonian Magazine
  • Integrated governance. GEIPAN’s steering committee structure, with civil aviation, meteorology, psychology, and national security seats, puts cross-domain expertise into one room. This is a best-practice many countries are still trying to replicate. Geipan
  • Methodological conservatism. GEIPAN’s D share (~2–3% over decades) is low compared to public expectations, reflecting strict criteria, case re-appraisals, and a willingness to downgrade D when new data warrant it. Geipan

Key Takeaways for Analysts 

  1. Most cases are explainable: A+B ≈ two-thirds across the archive; C ≈ one-third are data-insufficient; D ≈ 2–3% persist after investigation. Treat C as missing information, not mystery. Geipan
  2. D1 vs D2 matters: D1s often hinge on credible testimony without sensors; D2s require strong consistency (multiple independent witnesses and/or instruments and/or traces). Assign higher analytic weight to D2s. Geipan
  3. Temporal bias: Recent years under-reflect true D share until complex cases complete investigations and get published (“dynamic stats” effect). geipan.fr
  4. Hoaxes are rare: focus analytic effort on misidentification patterns (astronomy, re-entries, drones, aircraft) and perception factors. Geipan

Practical: How to Engage with GEIPAN

Methodology Deep-Dive 

Cognitive Interview & On-Site Work

For D1/D2 candidates, GEIPAN conducts cognitive interviews, scene walk-backs, and expert reviews. The office explicitly ties D classification to weirdness > 0.5 and appropriate consistency, and requires on-site work for D1/D2. Reclassification is always possible. Geipan

Imaging/Video Authentication

GEIPAN’s standard is IPACO® (co-funded with CNES). It combines metadata/forensic checks with photogrammetry and contextual tools. This makes France unusual for having a standardized, state-vetted imaging workflow for UAP. Geipan

Steering & Security Interfaces

GEIPAN can request radar checks via the Air & Space Force; Gendarmerie and Air Force representatives sit on the steering committee and can recommend actions. That interface enables multi-sensor cross-checks without making GEIPAN a security/intelligence shop. Geipan

Implications

  1. Scientific: A small but persistent tail (D) remains after rigorous investigation. Distinguishing D1 from D2 is crucial, the latter are the cases where multi-witness/sensor convergence suggests an unknown that merits prioritized study and, ideally, instrumented follow-up. Geipan
  2. Policy: The COPEIPAN model provides a template: define transparent roles, publish cases, and maintain a standing forum where aviation safety, meteorology, psychology, defense, and space meet routinely. Countries building national UAP processes could adopt similar governance to avoid stovepipes. Geipan
  3. Public trust: Publishing primary case files (not just summaries) is the single strongest lever France uses to defuse speculation and support independent reanalysis. The 2007 archive release set that norm. Smithsonian Magazine

Recommendations

For GEIPAN (and partners):

  • Publish a programmatic D2 registry with standardized metadata (sensor presence, independent witness count, trace/imagery flags) to make D2 machine-queryable across years. (Current widget obscures D2 by rounding; the IPACO post reaffirms D2’s importance.) Geipan
  • Open an analyst API for the case database (JSON/CSV endpoints), exposing classification, timestamps, geo-fuzzed coordinates, and phenomena tags to enable external replication and time-series studies. Geipan
  • Standardize sensor tasking: publish a playbook for when GEIPAN requests radar checks, ADS-B/Mode-S pulls, and astronomical cross-checks, including audit trails. Geipan
  • Expand imagery ground truth: continue co-development of IPACO®, add public test sets (known hoaxes, lens artifacts, drone light patterns) to benchmark community tools. Geipan
  • Case lifecycle dashboards: add a time-to-classification metric by year and class to correct the “recent years have fewer D” perception caused by investigation lag. geipan.fr

For other national programs (lessons learned):

  • Emulate France’s transparency: publish anonymized primary files with clear classification logic (including “weirdness” & “consistency” definitions).
  • Build a steering committee with military + civil + scientific representatives, explicitly chartered to recommend, not control, case outcomes, preserve analytic independence. Geipan
  • Prioritize witness experience quality: adopt cognitive interviews, user-friendly online forms, and privacy-by-design retention policies (with declared durations, as France does). geipan.fr

FAQ-Style Quick Answers

  • How many cases are really “unknown”?
    About 2–3% after investigation, depending on epoch and re-appraisals. Geipan
  • Does GEIPAN consider non-human craft?
    GEIPAN says it cannot exclude unknown origins in principle, but has no proof of extraterrestrial craft to date. Geipan
  • Can pilots report to GEIPAN?
    Yes, pilots, ATC, police, gendarmerie, and the general public all appear in the database. GEIPAN sometimes posts pattern explainers tailored to aviation (e.g., cockpit “flashes”). Geipan

Final Word

GEIPAN is the oldest continuously operating government UAP office with a public, case-level archive. Its data-first approach demonstrates that most sightings yield to mundane explanations, yet a small, resilient set of D1/D2 cases remains. Rather than closing the book, GEIPAN shows how to keep it open responsibly: standardized intake, multi-domain analysis, measured language, and public files that let the evidence speak. This is the template to professionalize UAP study worldwide. Geipan

References

  1. CNES – GEIPAN project page (EN): https://cnes.fr/en/projects/geipan.
  2. GEIPAN – Statistics (live): https://www.cnes-geipan.fr/en/stats.
  3. GEIPAN – Classification & methodology (EN): https://www.cnes-geipan.fr/en/node/58787.
  4. GEIPAN – Missions / Methods / Results (EN): https://www.cnes-geipan.fr/en/missions-methodes-et-resultats.
  5. GEIPAN – Homepage & “Actualités” (FR/EN) incl. IPACO®: https://www.cnes-geipan.fr/; IPACO post: https://www.cnes-geipan.fr/fr/node/61025.
  6. Gendarmerie Nationale – How PAN reports are handled (FR): https://www.gendarmerie.interieur.gouv.fr/gendinfo/actualites/2022/comment-la-gendarmerie-prend-elle-en-compte-les-etrangetes-dans-le-ciel.
  7. GEIPAN – “National Defence & GEIPAN” FAQ (radar checks, committee presence): https://www.cnes-geipan.fr/.
  8. GEIPAN – Witness guidance & forms (FR): https://www.cnes-geipan.fr/fr/qu-ai-je-vu/etape-2 + form PDFs.
  9. GEIPAN – Case search (EN): https://www.cnes-geipan.fr/en/search/cas. geipan.fr
  10. GEIPAN – Trans-en-Provence case page (FR): https://www.cnes-geipan.fr/fr/cas/1981-01-00849.
  11. France-Science (Embassy/Office for Science & Technology) CAIPAN II (2022): https://france-science.com/caipan-ii-colloque-international-consacre-aux-pan-organise-par-le-geipan-a-toulouse/.

SEO Keywords

GEIPAN, CNES UAP, France UAP office, PAN France, GEIPAN statistics, D1 D2 UAP France, French Gendarmerie UAP, IPACO image authentication, COPEIPAN, Trans-en-Provence case, report a UAP France, CAIPAN conference, UAP data France, GEIPAN methodology

Was this article helpful?
Yes No

Related Articles

Want to contribute?

Are you interested in working with us?
Request to Join