The Belgian wave brought together reports of unusual night-time lights and apparent structures, police observations and an Air Force effort to identify possible intruders. These were different kinds of evidence. Civilian reports occurred on separate dates and under different viewing conditions; Wilfried De Brouwer described the military response from his position in Air Staff operations. The F-16 interception of 30–31 March 1990 neither produced a pilot visual identification nor established that the radar contacts were the objects reported months earlier. Its early extraordinary-performance interpretation was subsequently qualified by technical analysis and by De Brouwer himself. The Petit-Rechain photograph has a separate provenance and a later admission of fabrication.
Residents and other civilian observers reporting separate events during the Belgian wave
Belgian civilian witnesses
Civilian reports and their reporting histories
The reports became widely known following the Eupen police observations of 29 November 1989. Civilian descriptions collected around this period included triangular light arrangements, central red lights, slow movement and apparent dark platforms. Those features were not universal. De Brouwer's 2010 retrospective also describes rectangular and trapezoidal forms, including a family travelling west of Liège and two factory workers at Basècles on 22 April 1990. These are accounts transmitted by the later author, rather than his own visual observations.
A more closely documented civilian example is the anonymous Mr J interviewed by SOBEPS investigator Auguste Meessen on 15 January 1990. Mr J described a silent lozenge-shaped object crossing above his Eupen home on 1 December 1989, followed by another observation on 10 January. His daughter arrived after the first object had receded; his wife did not see it. Meessen records that the drawing's vertical edge came from the second sighting. The published illustration therefore combines two occasions rather than reproducing every detail seen on the first evening.
Mr J estimated the first object's span using buildings about 35–40 metres apart, but could not determine its altitude. The account preserves useful local reference points without supplying a surveyed distance to the object. His estimate of a very large structure remains an interpretation of the reported view. Meessen's confidence in the witness's sincerity is an investigator's judgement, distinct from independent measurement.
The Petit-Rechain slide was presented as an April 1990 observation and later became a visual emblem of the wave. De Brouwer defended its authenticity in his 2010 chapter. On 26 July 2011, RTL broadcast an admission by its maker that it was a staged model photograph; VRT reported the admission the next day. The image cannot corroborate the earlier civilian accounts. Their existence also does not restore its authenticity.
Evidence review
The collective entry preserves reports with unequal detail and unequal proximity to the event. Meessen's published interview permits a more specific assessment than an aggregate witness count: it identifies the interview date, differing family viewing access and details added from a subsequent sighting. A total assembled across months cannot establish that everyone saw the same object, or that each report was independent of previous publicity.
The same investigator provides an important limit on instrumental corroboration. In his near-event article, Meessen reports finding no reliable ground-radar traces associated with the visual accounts he examined, while interpreting a separate class of slow radar echoes as atmospheric. His proposed reflective-platform explanation for non-detection is a hypothesis about the reported objects, not a measured radar signature from them.
Wim Van Utrecht's critical reconstruction proposes different stimuli for different reports, including aircraft, astronomical lights and helicopters. He also questions whether later elaborations of the Eupen account preserve the initial descriptions. These are specific investigative arguments, but they do not identify a flight corresponding to Mr J's two observations. De Brouwer's rejection of a general helicopter explanation similarly supplies no calibrated range or traffic match for that civilian example. The disagreement requires examination of individual viewing geometry and source history, rather than a verdict transferred from the entire wave.
Statement and reporting history
29 November 1989: Eupen police reports and civilian observations become the principal early cluster.
1 December 1989: Mr J's reported first Eupen observation.
10 January 1990: Mr J's reported second observation.
15 January 1990: Meessen interviews Mr J; the published drawing includes a detail from the second sighting.
April 1990: claimed date of the later Petit-Rechain slide, separate from the opening reports.
November 1990 / May–June 1991: Meessen's investigator account appears in Inforespace and in an English IUR translation.
2010: De Brouwer publishes a retrospective sample of witness reports and endorses the photograph.
26–27 July 2011: RTL broadcasts the fabrication admission; VRT reports it.
Ten-page scan of a translated November 1990 Inforespace article. Relevant PDF 1–6 read in extraction; scans 1, 3–6 inspected. Printed 8–9/PDF 5–6 contain Mr J's interview and combined drawing; PDF 3 records the negative ground-radar correlation finding. Complete original interview notes were not inspected.
Relevant chapter read in extracted text, PDF 20–30 of this edition. First-person operational retrospective includes relayed witness descriptions, photo endorsement and explicit March-interception limitation. The book predates the 2011 photo admission; original image-analysis reports were not inspected.
Both web instalments read, including revision dated May 2007 and their named references. Critical reconstruction and specific conventional proposals belong to Van Utrecht and the cited investigators; original initial Eupen interview and all proposed flight matches were not independently recovered.
Original broadcaster page and 13.00 publication date checked. Embedded report was not played or transcribed during this review; admission details checked against the VRT report of 27 July 2011 and later investigator account.
Broadcaster text published on 27 July 2011 obtained through indexed retrieval; direct page access was denied by robots. Reports the photographer's polystyrene-model admission.
Complete web addendum read. Describes the author's model experiments and correspondence with the photographer, while noting disputed early distribution accounts. Original slide, correspondence and other exposures were not inspected.
Chief of the Belgian Air Staff's Operations Section during the response; later retrospective author
Wilfried De Brouwer
Operational decisions and the evolving radar assessment
The cover attribution accompanying the July 1990 summary identifies De Brouwer as Chief of the Operations Section of the Belgian Air Staff. His book contribution, preserved in a UK Ministry of Defence file, describes attending the SOBEPS press conference on 18 December 1989 in that capacity. This establishes his operational role without implying that he personally witnessed the reported triangular objects or flew either interceptor.
De Brouwer describes an identification procedure using existing radar stations and F-16 readiness aircraft. Police confirmation and radar detection were required before a scramble. His archived account emphasises the delays introduced by telephone calls, patrol attendance, brigade radio and controller communication. It also records an earlier mission in which fighters identified projected lights over the Hasselt area. The military response therefore included both unsuccessful searches and a conventional identification.
The summary associated with his 11 July 1990 briefing records nine interception attempts after the two F-16s took off at 00.05 on 31 March. It reports six lock-ons, displayed speeds between 150 and 1,010 knots and three occasions of simultaneous locks. Neither pilot obtained visual contact. Its exclusion of several conventional explanations was the Air Force's reported assessment at that stage, dependent on how the radar values and ground reports were interpreted.
The archived book contribution is already more qualified than a simple radar-confirmation claim. It discusses spurious atmospheric echoes, asks why no definite correlation was found between the Ramillies observations and the fighter recordings, and leaves aircraft versus exceptional electromagnetic interference open. It nevertheless treats the wider testimony and reported radar agreement as grounds for considering unauthorised activity. This combines an operational account with an interpretation of evidence supplied by other observers.
In his 2010 chapter, De Brouwer says the technical analysis was unfinished at the July briefing and that only one aircraft camera provided a satisfactory radar recording. Consequently, a recorded comparison between the two fighters was unavailable. He describes the Air Force's conclusion as insufficient evidence to prove real craft on that occasion, while continuing to defend the broader civilian and police reports. His confidence in those reports is his retrospective position; it is not a replacement for the missing instrument comparison.
Evidence review
De Brouwer describes a response he helped administer: the decision to investigate, conditions for launching aircraft, communication difficulties and what staff could conclude. His position also explains his access to witnesses and radar summaries. It does not make every reported range, shape or display value a direct observation by him.
The early July summary, the archived book account and the 2010 chapter express different degrees of confidence about radar corroboration. The early briefing's reported simultaneous locks are not equivalent to two surviving, independently comparable recordings. De Brouwer's later acknowledgement of this distinction materially limits claims that the F-16s measured one object performing extraordinary manoeuvres.
His 2010 endorsement of the Petit-Rechain image predates the 2011 fabrication admission. The sequence matters when evaluating his photographic argument. His advocacy of further investigation remains an attributable policy position, while the image supplies no reliable support for the objects or propulsion mechanism described in that chapter.
Statement and reporting history
18 December 1989: De Brouwer attends a SOBEPS press conference as Air Staff operations chief.
30–31 March 1990: ground reports and radar contacts prompt two F-16s to launch.
11 July 1990: De Brouwer presents the military summary while the technical interpretation remains under examination.
Early 1990s: his book contribution discusses ground-radar filtering, communication delays and uncertainty over source correlation; an English version survives in DEFE 24/1960.
2010: first-person retrospective distinguishes insufficient March-interception evidence from continued confidence in the broader wave.
18 June 2025: Belgian Defence supplies a redacted copy of the historical Salmon technical report.
February 2026: Marc Hallet adds a separate preface to the public report compilation.
Complete short transcription read, including Lamin's cover attribution, reported locks, no pilot visual contact and source-owned exclusions. Original signed summary scan not inspected; reported observations are not raw instrument data.
English contribution's five pages at PDF 321–325 visually read in the 487-page archive mirror. Contains ground-radar filtering, communication chain, earlier conventional identification and unresolved March correlation. This is not the July press-summary document; surrounding file was not comprehensively inspected.
Relevant chapter read in extracted text, PDF 20–30 of this edition. First-person operational retrospective includes relayed witness descriptions, photo endorsement and explicit March-interception limitation. The book predates the 2011 photo admission; original image-analysis reports were not inspected.
173-page compilation. Selected original scans visually inspected: Defence release PDF 5–6; contents/introduction 7–10; reconstruction reassessment 47–48; Kalman discussion and episode reconstructions 50–64; conclusion 65–67. Historical report completion date not established from those scans. Redactions affect technical detail. Hallet's February 2026 preface is separately authored.
Eight hosted original report scan images visually read and compared with the French and English transcriptions. Original scan 2 begins the first call at 22.50, unlike the 23.00 transcription. Scans 6–7 contain the inversion inconsistency and one-return/eight-points limit. Listed operational annexes were not inspected. The separate five-page English PDF is a modern re-typeset reproduction.
Original broadcaster page and 13.00 publication date checked. Embedded report was not played or transcribed during this review; admission details checked against the VRT report of 27 July 2011 and later investigator account.
What the radar investigation adds to the witness record
Major Lambrechts's preliminary report treats the March observations as correlated across military sensors and gives considerable weight to the gendarmes' objectivity. It also records eight visually reported points with only one radar contact at a time and no pilot visual contact. These qualifications leave the association between lights and returns less secure than its conclusion suggests. Its general-information section reports slight inversions, whereas a later exclusion argument denies an inversion during the radar observations; that internal difference cannot establish the absence of propagation effects.
The technical report prepared under Salmon's direction revisits the aircraft displays rather than accepting their parameters as direct, unprocessed measurements. Its initial reconstruction is explicitly revised after recognition that the inputs already incorporate a Kalman filter: a calculation that predicts target position between usable returns and updates that prediction when new detections arrive. The report explains how resumed detection can create abrupt display changes, how extrapolation can extend a track after loss of a target, and how a lock can switch to a different source.
The report treats interceptions 3, 8 and 10 as probably following an F-16, and regards portions of other apparently remarkable tracks as fictitious or affected by source switching. These are the report's numbered display episodes, not the nine operational attempts in the July summary. It expressly labels its interception reconstructions speculative and says insufficient precision and knowledge of the filter prevent deeper analysis. The technical argument substantially weakens the inference of extreme physical acceleration without demonstrating a unique identified cause for every return.
Its final conclusion also says the study cannot link the fighter-radar observations to the visual testimony on the available evidence. That limit is narrower than Marc Hallet's February 2026 preface, which argues categorically against an extraterrestrial source. The redacted Defence release letter dated 18 June 2025 establishes the disclosure history. It does not turn the historical report into a new 2025 investigation, and the archive index's 1990 date does not establish its completion date.
These findings change the relationship between the two witness panels. The civilian accounts retain their reported local observations and individual histories; the military source establishes an attempt to investigate them. Neither stream automatically verifies the other. The absence of a pilot visual identification, the missing second satisfactory recording and the demonstrated importance of processing make it inappropriate to use the March radar episode as an instrumental measurement of the entire wave.
The photograph presents a different evidential problem. Its maker's fabrication admission directly undermines that image's provenance, whereas the radar investigation questions the inference from processed display values to a physical trajectory. Both remove commonly repeated forms of corroboration. Neither supplies an identification of every earlier light or apparent structure reported by civilians. The surviving question concerns particular observations and their available records, rather than one craft inferred from a national report total.
Key dates
29 November 1989
Eupen-area police and civilian reports form the principal early cluster, preceding the Petit-Rechain image.
1 December 1989
Mr J reports a lozenge-shaped object above his home; Meessen later documents the civilian account.
18 December 1989
De Brouwer attends the SOBEPS press conference as Air Staff operations chief.
15 January 1990
Meessen interviews Mr J after the witness's second reported observation on 10 January.
30–31 March 1990
Reports and ground-radar contacts prompt the F-16 scramble; neither pilot visually identifies the suspected source.
11 July 1990
The military summary reports rapid display changes and provisional exclusions; technical interpretation continues.
May–June 1991
IUR publishes the translated Meessen article first issued in November 1990, including his negative ground-radar correlation finding.
2010
De Brouwer's retrospective acknowledges insufficient evidence for real craft during the March interception while defending the wider reports.
26 July 2011
RTL broadcasts the Petit-Rechain maker's fabrication admission.
18 June 2025
Belgian Defence releases a redacted copy of the historical Salmon technical report.
February 2026
Hallet's separately authored preface accompanies the public report compilation.
Ten-page scan of a translated November 1990 Inforespace article. Relevant PDF 1–6 read in extraction; scans 1, 3–6 inspected. Printed 8–9/PDF 5–6 contain Mr J's interview and combined drawing; PDF 3 records the negative ground-radar correlation finding. Complete original interview notes were not inspected.
Complete short transcription read, including Lamin's cover attribution, reported locks, no pilot visual contact and source-owned exclusions. Original signed summary scan not inspected; reported observations are not raw instrument data.
English contribution's five pages at PDF 321–325 visually read in the 487-page archive mirror. Contains ground-radar filtering, communication chain, earlier conventional identification and unresolved March correlation. This is not the July press-summary document; surrounding file was not comprehensively inspected.
Relevant chapter read in extracted text, PDF 20–30 of this edition. First-person operational retrospective includes relayed witness descriptions, photo endorsement and explicit March-interception limitation. The book predates the 2011 photo admission; original image-analysis reports were not inspected.
Eight hosted original report scan images visually read and compared with the French and English transcriptions. Original scan 2 begins the first call at 22.50, unlike the 23.00 transcription. Scans 6–7 contain the inversion inconsistency and one-return/eight-points limit. Listed operational annexes were not inspected. The separate five-page English PDF is a modern re-typeset reproduction.
173-page compilation. Selected original scans visually inspected: Defence release PDF 5–6; contents/introduction 7–10; reconstruction reassessment 47–48; Kalman discussion and episode reconstructions 50–64; conclusion 65–67. Historical report completion date not established from those scans. Redactions affect technical detail. Hallet's February 2026 preface is separately authored.
Both web instalments read, including revision dated May 2007 and their named references. Critical reconstruction and specific conventional proposals belong to Van Utrecht and the cited investigators; original initial Eupen interview and all proposed flight matches were not independently recovered.
Original broadcaster page and 13.00 publication date checked. Embedded report was not played or transcribed during this review; admission details checked against the VRT report of 27 July 2011 and later investigator account.
Broadcaster text published on 27 July 2011 obtained through indexed retrieval; direct page access was denied by robots. Reports the photographer's polystyrene-model admission.
Complete web addendum read. Describes the author's model experiments and correspondence with the photographer, while noting disputed early distribution accounts. Original slide, correspondence and other exposures were not inspected.