An official report into the fatal incident on September 12 last year, when a gas balloon with two persons on board participating in the traditional Coupe Gordon Bennett international balloon race was intercepted and shot down by a Belarussian military helicopter, has concluded that the helicopter crew had been led to believe that they were shooting at an unmanned weather balloon. A sequence of administrative errors and failures of communication on the part of the Belarussian civil aviation and military authorities had, according to the report, led to this false interpretation.
The report also suggests that the balloon’s pilots, Alan Fraenckel and John Stuart- Jervis of the Virgin Islands, may have been asleep or resting at the time of the interception; that they may not have had access to the necessary radio frequencies to contact air traffic control in Belarus; and that their balloon may not have been carrying the necessary identification markings.
The Federation Aeronautique Internationale, in a detailed and closely argued analysis of evidence, not all of which was taken into account by the official investigating team, has categorically rejected all three of these allegations.
Log-entries, radio conversations with other competitors, the fact that the emergency frequency of 121.5 was selected on the balloon’s radio, and barograph evidence all indicate that the crew-members were awake and fully alert when the interception took place. Frequency information was distributed before take-off to all competing crews by the organizers at the Swiss launch site, and the crew carried a GPS with database software, including frequency information. Furthermore, information from the balloon manufacturer and witnesses at the launch site prove beyond doubt that the balloon was carrying all the identification markings required by ICAO.
The Federation Aeronautique Internationale (FAI) has decided to request that the official investigation be reopened to examine all the evidence identified by the FAI which calls into question the report’s conclusions. It is working in conjunction with ICAO to ensure that international interception procedures take into account the special characteristics of balloons. And it has instructed international legal counsel to determine what follow-up action can be taken to prevent repetition of this tragic event and to obtain compensation.
Statement by Federation Aeronautic Internationale (FAI) on the Official Report into the Belarus Balloon Incident
Background
1. On 12 September 1995, a gas balloon (D-CARIBBEAN) participating in the historic Coupe Gordon Bennett international balloon race was shot down by an attack helicopter of the Belarus air defense forces with the loss of the lives of the two pilots on board, Alan Fraenckel and John Stuart-Jervis, American citizens representing the Virgin Islands Aero Club. An official investigation into the circumstances surrounding the incident was carried out under the auspices of the Flight Safety Commission of the Interstate Aviation Committee of the Commonwealth of Independent States. The investigating team included observers from the National Transportation Safety Board (NTSB) and the Luftfahrt-Bundesamt (LBA), the agencies responsible for accident investigation respectively in the USA and in Germany (where the balloon was built and registered).
2. The English language version of the report, translated from the Russian by the US government, reached FAI’s hands on 4 June 1996. At the direction of the FAI President, the FAI 1st Vice President and the President of the FAI Ballooning Commission, assisted by professional advisers on ballooning operations, military interception procedures and aviation medicine, conducted a detailed evaluation of the report.
3. FAI welcomes the fact that the investigators of the Interstate Aviation Committee went about the task of investigating the incident with the primary aim of preventing a re- occurrence of this tragic event, and notes with satisfaction the readiness of the Belarussian authorities to allow publication and analysis of data from the attack helicopter’s flight recorder. FAI also notes, and fully supports, the view of the NTSB’s accredited representative on the investigating team that the appropriate and sound recommendations contained in the report concerning interception procedures should be directed explicitly at the Belarus military authorities, as well as the civil aviation authorities.
4. FAI acknowledges the good faith of the investigators, and would like to place on record its appreciation for the detailed work they carried out. We have noted with interest their description of the errors of administration and communication that occurred in the Belarus civil aviation and military authorities, and which they consider to have been contributory causes of the incident.
FAI’s Comments
5. We nevertheless find it essential to point out a number of important omissions, inconsistencies and inaccuracies in the report, and specifically in the section dealing with events in the period immediately before the shoot-down:
a. Two other balloons participating in the Gordon Bennett race entered Belarus airspace at around the same time as D-CARIBBEAN. The investigators interviewed the pilot of one (D Levin - N69RW), who landed without interference. However, no mention is made in the report of the other (M Wallace - N1636S), who, a short time after the shooting-down of D-CARIBBEAN, was himself intercepted and forced to land at gun- point by a military helicopter. Mr. Wallace states that he was in frequent radio contact with the crew of D-CARIBBEAN until shortly before the shoot-down. The frequency used by Mr. Wallace was 154.515, a frequency selection which the investigators found difficult to explain on D-CARIBBEAN’s radio. The hypothesis advanced by the investigators that Mr. Fraenckel and Mr. Stuart-Jervis might have been asleep or in a state of extreme fatigue at the time of the shoot-down is therefore untenable, even if medically plausible. Not only did they make two log-entries in the 75 minutes preceding the shoot-down, but also Mr. Wallace states that he spoke to them on 154.515 during this same period and found them fully awake and alert, and showing no signs of stress. Indeed, both crews were jubilant at the imminent prospect of winning, or at least being well-placed, in the race.
b. No analysis of the balloon barograph trace is included in the report, nor is altitude information from this trace correlated with time/altitude data derived from the record of the flight parameters of the Mi-24 attack helicopter. The balloon’s barograph trace indicates that, before the near vertical plunge to earth resulting from the destruction of the envelope, the balloon had entered a high-rate descent of between -2.5 and -4.5 m/s. This is entirely consistent with the crew’s having initiated an emergency descent on seeing the apparently hostile intent of the helicopter. It is not consistent with the rate of descent to be expected as the envelope was in the process of disintegration after being hit. The barograph therefore provides further evidence that the crew were not only awake and alert, but also took action in the face of imminent danger. We are further concerned about certain unexplained features of the barograph trace (of which FAI has seen only a photocopy). There is no explanation in the report of why the trace suddenly thickens substantially some 15 minutes prior to the plunge, nor of why the vertical plunge appears to continue below the zero meter baseline to -1100 meters, nor of why there appears to be slowly rising trace after what is presumed to be the impact.
c. The report makes no mention of the emergency locator transmitter (ELT) which was carried by the crew of D-CARIBBEAN. It would have been of great interest to know whether the ELT had been activated. Mr. Wallace states that as soon as he was intercepted by the Belarus attack helicopter and initiated a descent, his ELT was activated. When the ELT is activated, no emergency voice transmissions on the international distress frequency of 121.5 are possible. This may explain why no such transmission was registered by satellite monitoring. In any event, the very fact that D- CARIBBEAN’s VHF radio was tuned to 121.5 provides further circumstantial evidence that the crew were fully operational and perceived an immediate threat to their safety. Balloon crews do not habitually fly with the international distress frequency selected, except in circumstances required by ICAO procedures. However, satellite monitoring should also have registered automatic transmissions emanating from Lets. Therefore, it would have been of great interest to know whether the satellite monitoring records show evidence of Mr. Wallace’s ELT transmissions. If they do not, their reliability is called seriously into question.
d. The report suggests that Fraenckel and Stuart-Jervis may not have had the necessary frequency for contacting Minsk ATC. During the official briefing on 8 September 1995, the organizers distributed to all participants and crews, (together with other material), a four-page document headed “Communications”, containing information taken from current Jeppesen IFR charts (Reference EE(H/L)1). This document included the latest officially promulgated frequencies for Minsk, Gomel and Grodno, as well as the radio station at Brest. Mr. Wallace confirmed that the Swiss competition organizers had indeed provided the necessary frequencies, including that of Minsk, and also stated that, in radio conversation with Mr. Fraenckel that morning, he had confirmed the Minsk frequency. Furthermore, the crew also had on board a GPS with Garmin 95 database software (not mentioned in the report), which comprises a wealth of aeronautical information for the whole of Europe, including Belarus. This software allows the user to dial up the frequency of the nearest ATC station at any time. The pilots of D-CARIBBEAN were both highly experienced air transport rated professional fixed-wing pilots (totaling between them 24,000 hours flying), entirely familiar with the procedural requirements for international flights.
e. There is no analysis or even mention in the report of power sources and levels of battery charge. D-CARIBBEAN was equipped with solar panels which made it quite independent of battery capacity considerations for radio transmissions.
We therefore have no explanation for the report’s finding that the crew of D- CARIBBEAN made no radio calls in Poland or Belarus, except that they may have unsuccessfully tried to contact Minsk (some 237 km from the crash site) before changing to 121.5. It goes without saying that an absence of radio calls is no justification for shooting down a civil aircraft.
f. There is no comparative analysis or explanation in the report of the different interception procedures used for the two balloons. If the Belarus air defense authorities (who, at the moment of shoot-down, according to the report, were under the mistaken belief that D-CARIBBEAN was an unmanned meteorological sounding balloon) were already aware of their fatal mistake when they intercepted Wallace’s balloon, it is difficult to understand why they apparently made no attempt to contact him by radio, and why they treated his balloon as a potentially hostile intruder. It would have been very illuminating to see in the report transcripts of the discussions that took place in the air defense headquarters during the period after D-CARIBBEAN had been shot down and before Wallace’s balloon was intercepted, and also of the conversations between the pilot of the helicopter that intercepted Wallace and his ground controller. It is noteworthy that this helicopter pilot, by approaching much closer, had no difficulty in identifying the balloon, and in communicating by hand signals that it should land.
g. The helicopter crew intercepting D-CARIBBEAN received orders to identify the balloon but to approach no closer than 50 meters. In fact, at the moment of first opening fire, the helicopter crew was, according to the report, still at a distance of about 1075 meters from the balloon. Crucially important, (a factor not mentioned in the report) the helicopter was also down-sun of the balloon throughout the maneuvers it made prior to opening fire. The sun’s azimuth (152 degrees) and elevation (38 degrees) at that date, time, latitude and longitude would probably have made it very difficult for the helicopter crew to distinguish visually any significant features of the balloon at a distance exceeding one kilometer. Certainly, the registration markings (which, as the manufacturer has confirmed, were of ICAO regulation size and displayed on both sides of the envelope as well as across the top, thus trebling, from 0.2 to 0.6, the probability quoted in the report of their being visible), would most likely have been invisible under those conditions, as would gesturing crew members. No explanation is provided in the report of why the crew did not inspect the balloon more closely and fly around it to the up-sun side before initiating the attack. If the crew considered they were dealing with a weather balloon, which presented no immediate threat or danger to air traffic, there seems no justification for their having failed to carry out a positive identification. The report does not mention whether the Mi-24 helicopter was equipped with video- recording equipment, as is usual for interceptor aircraft in many countries. If it was so equipped, the video recording of the interception would of course have yielded invaluable evidence about the aspect presented by D-CARIBBEAN from the viewpoint of the helicopter crew.
h. The fact that the helicopter crew opened fire from a distance with minimal precautions is certainly consistent with a pre-established belief that the balloon was unmanned. However, in the transcript of the conversation between the combat control officer (CCO) and the helicopter crew (HC) after the balloon has been hit and is plummeting to earth, the following exchange takes place:
CCO: “On the basis of visual observation, does the gondola have any equipment, components?”
HC: “Can’t tell what was in the gondola, maybe equipment, but no people”
The report does not pick up on the lack of logic of this exchange. If the crew could not tell what was in the gondola, it is difficult to explain why they should categorically state that there were no people. If they had been led to believe they were shooting down a weather balloon, there would have been no reason to mention people.
i. The report’s reconstruction of the relative positions of the balloon and intercepting helicopter during the interception phase is apparently based on the helicopter crewman’s statement, just before opening fire for the second time, that the range from the target was 600 meters. No explanation is given of how this range was calculated. If the helicopter was equipped with a laser range-finding device, and the range was therefore dependably accurate, the crew should immediately have realized that the target was far too large for a weather balloon. If, on the other hand, there was no range-finder, the calculation must presumably have been based on the supposed size of the “weather” balloon and the angle subtended by the target as perceived by the helicopter crew. The total size of D-CARIBBEAN, as confirmed by its German manufacturer, Michael Woerner, was 25 meters overall height, and 12.6 meters envelope diameter, that is some ten times the size of the average weather balloon (the report gives no details of the actual size of the Polish weather balloon). It is therefore possible that the helicopter was in fact considerably further away from D-CARIBBEAN than its crew imagined. Grossly underestimating the range could account for the fact that a professional attack-helicopter gunner missed a very large, slow-moving target by (according to him) 100 meters to the right, during a burst of fire extending over 9 seconds. (Another plausible explanation — apart from professional incompetence — is the possibility that the first volleys of fire were designed to miss and were in fact warning shots. The latter explanation is of course entirely inconsistent with the notion that the crew thought they were dealing with an unmanned weather balloon.)
j. The report makes no mention of whether recognition training is given to Belarussian military combat pilots. The gross disparity in size between a manned gas balloon and a weather balloon, allied to the different shape of the suspended loads in each case, should have been sufficient to create doubt in the mind of a trained combat pilot about whether he was looking at a weather balloon.
k. The report states that D-CARIBBEAN did not meet the requirements of the race organizers since “the gondola did not carry the race participant’s numeric symbol or identifying plaque.” FAI categorically rejects this finding. D-CARIBBEAN was displaying the two banners required by the race rules at the moment of launch, and this was verified by the International Jury. The banners were attached with wire ties threaded through the wicker-work of the basket. It is just conceivable that the banners were ripped off during the plunge to earth of the balloon, but there is no plausible reason why the crew should have removed the banners in flight. In any case, this is a feat that would have been physically extremely difficult, if not impossible, to achieve from within the basket. As for the national flag of the Virgin Islands, this was securely fixed high on the net rigging, at least 10 feet above the load ring, entirely out of reach of the pilots. The flag was made of nylon, and could therefore have burned during the fire. The flag found in the basket may have been a spare. Alternatively someone at the crash site may have thrown the flag into the basket. There is no way in which the crew could have removed the flag from its attachment point and put it in the basket. As previously pointed out, the display of banners and flag was in any case of academic interest, since the helicopter crew made no attempt to maneuver into a position from which these could have been seen and recognized.
l. The authors of the report state quite correctly that there are some requirements in the ICAO Interception Guide (Doc. 9433-AN/926) which are inapplicable to free-flying balloons (e.g. a balloon is unable to comply with an instruction to change direction in the horizontal plane). However, it is not necessarily correct to say that “the crew of a manned balloon cannot ...give signals by blinking the navigation lights... and(or) turning on the landing lights.” D-CARIBBEAN was equipped with a red/white strobe light for night flying, and with a powerful beam light for night landings. Either of these could have been activated by the crew if the interceptor had established radio contact on the distress frequency and made a request to this effect.
Conclusions
6. In the light of the evidence outlined above, FAI is persuaded that the Belarus incident report has serious deficiencies, and specifically that:
a. Contrary to the suggestion in the report that they might have been asleep or incapacitated, the pilots of D-CARIBBEAN were in fact awake and capable of responding at the time of the interception. If no response was noted by the helicopter crew, his was most probably because of the helicopter’s distance (down-sun) from the balloon, the fact that no attempt was made to establish contact on the international distress frequency, and the extremely short time interval that elapsed between the helicopter’s first arriving in the vicinity of the balloon and then shooting it down.
b. The balloon pilots selected the international distress frequency (and may possibly have activated the emergency locator transmitter) when they saw the attack helicopter approaching. They also appear to have initiated a maximum-rate descent.
c. Contrary to the suggestion in the official report that they might have been lacking in essential aeronautical information, such as frequencies, the pilots of D-CARIBBEAN had access to the frequency for Minsk ATC and all other aeronautical information required for penetration of Belarus airspace, from he race organizers and from other sources.
d. Notwithstanding the doubts expressed in the report, the manufacturer has confirmed that D-CARIBBEAN was displaying registration markings of ICAO regulation size (50 centimeters high) in 3 separate places: on both sides of the envelope, and across the top of the envelope. In addition, the balloon carried he regulation banners on opposite sides of the gondola and he national flag in the balloon rigging.
7. We further conclude that, whatever the sequence of events might have been that led the Belarus air defense authorities to confuse D-CARIBBEAN with a Polish weather balloon, these authorities were culpably negligent in their failure to comply with their own and with international regulations on interception procedures, and had no justification whatsoever for shooting down the balloon.
Action by FAI
8. FAI is conducting a review of ICAO Doc. 9433-AN 1926, with a view to drawing up recommendations relating to the interception of balloons and other unpowered aircraft. The ICAO has agreed to take FAI’s recommendations carefully into account when conducting the review of these procedures recommended by the Interstate Aviation Committee. The FAI’s permanent representative at ICAO Headquarters in Montreal has had preliminary discussions on the Belarus incident report with Mr. Assad Kotaite, President of ICAO, and is scheduled to meet him again in the near future for more detailed discussions.
9. All FAI Member organizations throughout the world are being made aware of the findings of the Belarus incident report and asked to consider whether any changes should be made to their procedures to prevent repetition of this kind of incident. The recommendations in the report (4.2.1 and 4.2.2) addressed to organizers of air sport competitions have been noted, but as stated in paragraphs 6c and 6d above, we believe that the report errs in identifying, as contributory causes of the present incident, a lack of aeronautical information held by the crew and inadequate markings on the balloon.
10. FAI will request that the Interstate Aviation Committee reopen its investigation to address thoroughly the issues outlined in this statement.
11. Finally, in view of the official report’s finding that the Belarus civilian and military authorities failed to comply with their own internal regulations on how requests for airspace penetration and flight plans should be processed, and with international conventions on procedures for interception of unidentified aircraft, FAI has had a meeting with, and instructed international legal counsel to establish what follow-up action could be taken in this case. On receipt of this legal advice, FAI will act in the best interests of all its Members world-wide, in fulfillment of the unanimously adopted resolution of the 88th FAI General Conference to “...undertake any action, including but not limited to the civil and criminal pursuit of any person and/or agency involved in this tragic event, that will contribute to: