The crew of Japan Air Lines cargo Flight 1628 reported unfamiliar lights over Alaska on 17 November 1986. Anchorage air-traffic controllers investigated the report during the flight, authorised manoeuvres and asked other aircraft to look. Captain Kenju Terauchi subsequently described two smaller craft and a much larger object. First Officer Takanori Tamefuji and Flight Engineer Yoshio Tsukuba supplied more limited visual accounts and their own descriptions of the aircraft's weather-radar display. The Federal Aviation Administration's later ground-radar assessment provides a separate evidential record. Comparing these sources requires retaining the differences between what each observer saw, what was relayed by radio and what investigators inferred from recorded data.
Captain and principal source of the detailed craft and large-object interpretation
Kenju Terauchi
The captain's account
Terauchi's written statement, received by the FAA on 2 January 1987 and translated by Sayoko Mimoto, describes aircraft-like lights to the left during the approach from northern Alaska. He initially considered military aircraft. He then reported two luminous forms suddenly appearing ahead, with light entering the cockpit and a feeling of warmth on his face. He interpreted their lights as exhaust arrangements and estimated a distance of 500–1,000 feet. These descriptions combine perceived appearance with his explanation of machinery and motion.
The narrative places a later pale light farther away and records Terauchi setting the colour weather radar to a 20-mile range. He described a green return seven or eight miles away in the same general direction. As the flight approached Fairbanks, he reported seeing a large silhouette behind and to the left. His 2 January FAA interview compares the larger object with twice an aircraft carrier. The interview also records uncertainty about departure direction and his belief that the object deliberately remained on the darker side of the aircraft.
The flight's contemporary radio transcript records unidentified traffic at approximately the aircraft's altitude and successive position estimates. At 02:30 UTC the flight describes something quite big, then requests descent to flight level 310, approximately 31,000 feet on the standard pressure setting. A right 360-degree turn follows at the controller's request. The flight later reports losing visual contact. The transcript identifies transmissions as JL1628 rather than assigning every call to Terauchi; the cockpit testimony indicates that the first officer also participated in the reporting.
Evidence review
FAA security notes from the post-flight interview already record erratic lights, an estimated object at least as large as a Boeing 747 and a claimed weather-radar indication. The more elaborate January account gives exhaust, warmth and large-silhouette interpretations in substantially greater detail. Investigator Jim Derry's initial summary says the crew saw lights without a craft, while Ronald Mickle's notes separately attribute size and joined-light impressions to Terauchi. These early summaries are investigator accounts of a joint interview; their different phrasing should remain visible when tracing the description's development.
The aircraft command role gave Terauchi substantial operational access, but the unknown lights supplied no established scale. His near-range estimates, the weather-radar distance and the later enormous size concern different phases and cannot be combined into one measured trajectory. No photograph resulted from his attempted camera use. Tsukuba's interview independently recalls handing him the camera and the failure to obtain an image.
The translated statement recounts climbing to 3,100 feet after the turn. The contemporaneous radio record instead documents descent from flight level 350 to 310 before the full turn. The narrative's reconstructed dialogue is therefore unsuitable as a substitute for the operational altitude sequence. Its interpretation of intentional concealment likewise remains Terauchi's inference from darkness and apparent relative position.
Statement and reporting history
17 November 1986, Alaska time: unfamiliar traffic is reported during the flight and the crew is interviewed after arrival.
18 November 1986, UTC: the preserved radio record covers the in-flight communications across the date boundary.
2 January 1987: the FAA receives the translated written account and conducts a further interview with Terauchi.
5 March 1987: the FAA releases its investigation materials and concludes that its systems did not confirm the reported event.
Mickle and Derry summaries at PDF pages 2–5, read as text; pages 2–4 visually inspected. Describes the 17 November debriefing. Derry's form has a conflicting year in its date field, while the narrative explicitly identifies the 1986 event; do not derive a new event date from that field. Joint summaries do not isolate every crew member's wording.
Complete 13-page English reproduction read, PDF pages 2–14; translation attributed to Sayoko Mimoto and received by FAA on 2 January 1987. Narrative contains reconstructed dialogue and altitude discrepancies compared with the radio transcript. Japanese original and translation alignment were not reviewed.
Selected account, radar and interpretation passages read, PDF pages 3–10, 12, 16–17 and 19; key pages 5 and 7 visually inspected. Interview includes JAL interpreter Frank Fujii and FAA's Sayoko Mimoto. Audio not auditioned; questions and interpreter contributions remain distinct from the captain's statements.
January 1987 certified transcript covering 18 November 1986, 02:14–02:59 UTC, read throughout. Selected timed passages visually inspected, including position mismatch, search coordination and JAL loss of sight before the C-130 check. File contains an early duplicated page. Transmissions identify the flight, not each cockpit speaker. Original tape not auditioned.
Complete interview text at PDF pages 2–6 read and visually inspected; original page order is 1, 3, 5, 2, 4, followed by a sketch. Interpreter Sayoko Mimoto. Explicitly distinguishes radar indication from initial lights and reliance on the captain during descent. Audio not auditioned. Name spelling varies across FAA records.
Release at PDF page 4 and scope statement at page 7 visually inspected; selected surrounding notes read. FAA could not confirm the November event and did not scientifically analyse celestial positions or viewing angles. Pages 5–6 discuss separate sightings, including the 11 January 1987 town-light explanation; that conclusion is not a finding on the November scene.
First Officer Takanori Tamefuji and Flight Engineer Yoshio Tsukuba, excluding Captain Terauchi
JAL flight crew
The first officer and flight engineer
In his 5 January 1987 interview, Tamefuji described the first lights as resembling oncoming traffic, with changing intensity and several colours. He repeatedly distinguished the lights from a visible complete body. He estimated his principal observation at roughly five to ten minutes. From the right-hand seat, his view of the later left-side phenomenon was obstructed by the captain and window pillars, while he was responsible for flying. He recalled captain and first officer noticing the initial lights at almost the same time and said he suggested contacting air-traffic control.
Tamefuji saw a green echo on the colour radar, at an indicated seven or eight miles, after the captain adjusted the instrument. He considered its appearance similar to aircraft echoes he had encountered previously. He was less definite about motion, saying that he could not determine its direction. His judgement about an aircraft-like return is thus a specific instrument interpretation, rather than confirmation of the larger object's outline or dimensions.
Tsukuba's 15 January interview describes clusters of white or amber lights ahead, resembling windows of a passenger aircraft, followed by much fainter lights to the left. He expressly separated the two appearances and said he could not describe a body. For the later lights he left open whether he was seeing a distant town or an unfamiliar object. During the altitude change he could not see the phenomenon and relied on the captain's statement that it remained present.
Tsukuba also saw a green, stream-like radar indication, which he estimated at about ten miles. He did not think it corresponded to the initial forward lights. His memory of the radar tilt was uncertain. He reported no magnetic disturbance or abnormality on his engineer's panel, and no internal-light reflection noticed in the window. These are his observations, without a saved radar image or optical reconstruction to establish the return's cause.
Evidence review
The two interviews support a shared observation of conspicuous lights, while giving different limits on the later phase. Tamefuji's active flying duties and restricted view explain why he could not reproduce all of Terauchi's description. Tsukuba's explicit reliance on the captain during descent is particularly consequential: it prevents that interval from becoming an additional independent observation that a target followed the manoeuvre.
Both supporting crewmen saw a weather-radar indication. Their testimony therefore gives the onboard display claim a broader basis than Terauchi alone. The estimated range and description differ, and Tsukuba rejects a direct connection with the initial lights. Their instrument impressions cannot by themselves join every visual phase to a single physical target.
The interviews were conducted through interpretation and include repeated questions, uncertain wording and non-verbal demonstrations. The first officer's detailed transcript preserves his own qualifications, whereas the initial joint summaries compress the three men into one account. Individual January testimony is essential for testing precisely which claims were shared.
Statement and reporting history
17 November 1986: the supporting crew members take part in the flight and post-flight debriefing.
5 January 1987: Peter Beckner interviews Tamefuji with interpreter Sayoko Mimoto.
15 January 1987: Beckner and Mimoto interview Tsukuba, preserving his distinction between initial and later lights and his limits during descent.
Mickle and Derry summaries at PDF pages 2–5, read as text; pages 2–4 visually inspected. Describes the 17 November debriefing. Derry's form has a conflicting year in its date field, while the narrative explicitly identifies the 1986 event; do not derive a new event date from that field. Joint summaries do not isolate every crew member's wording.
Complete interview text read, PDF pages 2–24; relevant account, radar and viewing-limit pages visually inspected. Interpreter Sayoko Mimoto. The record distinguishes lights from a resolved body and identifies restricted later viewing. Original audio not auditioned.
Complete interview text at PDF pages 2–6 read and visually inspected; original page order is 1, 3, 5, 2, 4, followed by a sketch. Interpreter Sayoko Mimoto. Explicitly distinguishes radar indication from initial lights and reliance on the captain during descent. Audio not auditioned. Name spelling varies across FAA records.
January 1987 certified transcript covering 18 November 1986, 02:14–02:59 UTC, read throughout. Selected timed passages visually inspected, including position mismatch, search coordination and JAL loss of sight before the C-130 check. File contains an early duplicated page. Transmissions identify the flight, not each cockpit speaker. Original tape not auditioned.
Anchorage controllers, supervisors and specialists who handled or reviewed the report; military radar information reached them through inter-agency communication
FAA radar personnel
The controller record and subsequent review
Carl E. Henley's statement dated 19 November 1986 records receiving the traffic query, checking military operations and seeing occasional primary radar indications. Samuel J. Rich subsequently described a display return near the reported position after reducing the displayed area. Their calls to the military Regional Operations Command Center sought both traffic identification and radar information. These personnel were observing display data and receiving crew reports; they had no direct visual view of the reported objects.
The contemporary transcript preserves disagreement about correlation. At 02:25:02 UTC the controller reports a hit about five miles behind the 747. The flight replies that its traffic is ahead-left, about eight miles away. Shortly afterwards the military operator describes a possible primary return while questioning whether it is erroneous; Anchorage requests continued monitoring. The discussion later includes loss of contact and renewed possible indications. At 02:51:49 Anchorage tells United Flight 69 that it has had only a few primary hits and never a good track.
Henley's 6 January clarification distinguishes tentative, radar-only track and other display symbols. He also states that his original description of traffic staying with the 747 through turns and descent came from the flight's reports. Supervisor Joseph Rollins says he never saw radar data conclusively establishing the reported traffic. Automation specialist Manfred Keller and quality-assurance specialist Anthony Wylie report finding no corresponding other-target information in the recorded data they examined.
Dennis R. Simantel's analysis dated 25 February examines approximately 61 minutes of recorded data. It attributes nearby separated primary and beacon indications to ordinary correlation behaviour between radar reflection and transponder reply. The Airway Facilities Division endorses the analysis on 27 February. The FAA's 5 March release adopts the split-return explanation and states that the event could not be confirmed. These are identified analytical findings, distinct from controllers' immediate reactions to a potentially urgent traffic report.
Evidence review
The controller documents establish active operational investigation and real-time uncertainty. The January clarification prevents an apparent claim of radar-confirmed following from being read beyond its source. A controller can responsibly act on uncertain traffic information while the later data examination finds a processing explanation for what appeared on the display.
The split-return analysis concerns the recorded ground-radar evidence examined by the FAA. It does not analyse the aircraft's unsaved weather-radar display or establish the source of every visual light. Military operators' spoken impressions are preserved in inter-agency dialogue, with intermittent contacts and expressed uncertainty. That dialogue supplies neither a continuous military track nor an independently verified outline of the large object.
Statement and reporting history
17–18 November 1986: Anchorage and military personnel exchange traffic information during the flight, recorded on the UTC date of 18 November.
19 November 1986: Henley signs his initial operational statement.
6–9 January 1987: clarification, controller, supervisor and data-review statements distinguish the different forms of access.
25–27 February 1987: Simantel completes his analysis and the Airway Facilities Division endorses it.
5 March 1987: the FAA publishes its technical explanation and closes the inquiry.
January 1987 certified transcript covering 18 November 1986, 02:14–02:59 UTC, read throughout. Selected timed passages visually inspected, including position mismatch, search coordination and JAL loss of sight before the C-130 check. File contains an early duplicated page. Transmissions identify the flight, not each cockpit speaker. Original tape not auditioned.
PDF pages 2–9 read; Henley, Rich, Rollins, Keller and Wylie statements visually inspected. Henley's 6 January clarification says following during manoeuvres was information received from JAL. Later pages repeat these statements. Aarnink and other forms contain date/time inconsistencies, so precise chronology follows the certified transcript.
All four substantive pages read; technical analysis at PDF page 5 visually inspected. Simantel analysis dated 25 February 1987, Morse endorsement 27 February, Steucke explanation and diagram 5 March. Concerns FAA recorded ground-radar data, not the aircraft's unsaved weather-radar display.
Release at PDF page 4 and scope statement at page 7 visually inspected; selected surrounding notes read. FAA could not confirm the November event and did not scientifically analyse celestial positions or viewing angles. Pages 5–6 discuss separate sightings, including the 11 January 1987 town-light explanation; that conclusion is not a finding on the November scene.
Comparing the evidence
The aircraft's report concerns the evening of 17 November in Alaska, while its preserved radio exchanges use 18 November UTC. Keeping these conventions explicit avoids turning one event into two different dates. The transcript documents the flight at flight level 350 and its descent to 310 before the controller-requested full turn. Later narrative dialogue changes that sequence and misstates the altitude. Operational chronology should therefore follow the timed transcript, with witness recollection used to describe perceived appearance.
Crew agreement is concentrated in the early lights and an onboard radar indication. The large silhouette, immense dimensions and interpretation of exhaust or purposeful concealment rest chiefly on Terauchi. Tamefuji had a restricted view of the later left side; Tsukuba was uncertain about the faint later lights and did not personally observe their persistence during descent. A shared cockpit created overlapping evidence, but the records preserve meaningful differences that a collective retelling would erase.
The FAA identified a specific processing explanation for the ground-radar indications it reviewed. Primary radar receives reflections from a target's surface; a beacon system receives a coded transponder reply. If their positions fail to correlate within a processing cell, one aircraft may produce apparently separate indications. Simantel's analysis treats the observed data as normal for the system and comparable with another aircraft in the area. This explanation has documented technical support for that radar evidence. It cannot simply be transferred to the different onboard weather-radar instrument.
Two outside aircraft provided a limited direct check. United Flight 69 identified the 747 and its contrail against the lighter sky but saw no accompanying aircraft. Its search occurred while JAL was still describing the traffic's position. The C-130, TOTEM71, investigated after JAL reported losing sight of the phenomenon and reported no additional lights. Their results constrain claims of a persistently visible accompanying craft, with the C-130's later timing offering a weaker test of what had been visible earlier.
Ordinary aircraft were the crew's initial working explanation. Tamefuji retained an aircraft-like interpretation of the lights and radar echo, although controllers found no identified traffic matching the report. For the faint later phase, Tsukuba explicitly considered distant town lights. Interior reflection was raised during his and Tamefuji's interviews and rejected by them; the records contain no controlled recreation establishing that explanation. These are source-specific alternatives, whose applicability differs between the early conspicuous pattern and later poorly resolved lights.
The FAA's closure was narrower than a comprehensive optical identification. In a 5 March statement, public-affairs officer Paul Steucke says the agency did not undertake scientific analysis of stars, planets or viewing angles, concentrating on its systems and aviation safety. A separate January sighting by Terauchi was attributed to town lights affected by ice crystals; that finding concerned another flight and cannot identify the November scene. The inspected November records consequently support a mixed evidential assessment: substantial visual testimony, a reported airborne-radar indication and a ground-radar explanation, with the physical source of the lights unresolved.
Key dates
17 November 1986, Alaska time
Flight 1628 reports unfamiliar traffic during its approach through Alaska. Controllers seek radar and visual checks, and FAA personnel interview the crew after arrival.
18 November 1986, 02:19 UTC
The preserved radio transcript records the first traffic query and a report of two apparent sources ahead. This is the same Alaska evening across the UTC date boundary.
18 November 1986, 02:32–02:37 UTC
The transcript records descent from flight level 350 to 310, followed by a controller-requested right 360-degree turn.
18 November 1986, 02:50–02:55 UTC
United Flight 69 identifies JAL but sees no accompanying aircraft. JAL then reports losing visual contact; the C-130 conducts its later check.
19 November 1986
Henley records the operational radar impressions and crew reports in an initial statement.
2 January 1987
The FAA receives Terauchi's translated written statement and conducts a detailed interview.
5–15 January 1987
Separate interviews with the supporting crew and personnel clarifications distinguish observation from relayed information.
25–27 February 1987
Simantel's recorded-data analysis attributes the indications to normal radar correlation behaviour; the Airway Facilities Division endorses it.
5 March 1987
The FAA releases its materials, adopts the ground-radar split-return explanation and closes its aviation-safety inquiry without identifying the visual lights.
Mickle and Derry summaries at PDF pages 2–5, read as text; pages 2–4 visually inspected. Describes the 17 November debriefing. Derry's form has a conflicting year in its date field, while the narrative explicitly identifies the 1986 event; do not derive a new event date from that field. Joint summaries do not isolate every crew member's wording.
Complete 13-page English reproduction read, PDF pages 2–14; translation attributed to Sayoko Mimoto and received by FAA on 2 January 1987. Narrative contains reconstructed dialogue and altitude discrepancies compared with the radio transcript. Japanese original and translation alignment were not reviewed.
Selected account, radar and interpretation passages read, PDF pages 3–10, 12, 16–17 and 19; key pages 5 and 7 visually inspected. Interview includes JAL interpreter Frank Fujii and FAA's Sayoko Mimoto. Audio not auditioned; questions and interpreter contributions remain distinct from the captain's statements.
Complete interview text read, PDF pages 2–24; relevant account, radar and viewing-limit pages visually inspected. Interpreter Sayoko Mimoto. The record distinguishes lights from a resolved body and identifies restricted later viewing. Original audio not auditioned.
Complete interview text at PDF pages 2–6 read and visually inspected; original page order is 1, 3, 5, 2, 4, followed by a sketch. Interpreter Sayoko Mimoto. Explicitly distinguishes radar indication from initial lights and reliance on the captain during descent. Audio not auditioned. Name spelling varies across FAA records.
January 1987 certified transcript covering 18 November 1986, 02:14–02:59 UTC, read throughout. Selected timed passages visually inspected, including position mismatch, search coordination and JAL loss of sight before the C-130 check. File contains an early duplicated page. Transmissions identify the flight, not each cockpit speaker. Original tape not auditioned.
PDF pages 2–9 read; Henley, Rich, Rollins, Keller and Wylie statements visually inspected. Henley's 6 January clarification says following during manoeuvres was information received from JAL. Later pages repeat these statements. Aarnink and other forms contain date/time inconsistencies, so precise chronology follows the certified transcript.
All four substantive pages read; technical analysis at PDF page 5 visually inspected. Simantel analysis dated 25 February 1987, Morse endorsement 27 February, Steucke explanation and diagram 5 March. Concerns FAA recorded ground-radar data, not the aircraft's unsaved weather-radar display.
Release at PDF page 4 and scope statement at page 7 visually inspected; selected surrounding notes read. FAA could not confirm the November event and did not scientifically analyse celestial positions or viewing angles. Pages 5–6 discuss separate sightings, including the 11 January 1987 town-light explanation; that conclusion is not a finding on the November scene.