At 9.22pm on the cloudy Sunday evening of 19th October 1986, the inhabitants of a hitherto obscure South African border village called Mbuzini, heard a loud impact.
It was caused by a Russian made Tupolev TU134-A airliner carrying the first President of the People's Republic of Mozambique, Samora Machel. The aircraft flew into a hillside a few hundred meters inside South African territory.
Of the 43 people aboard, all but 6 of the passengers and one of the aircrew perished. The accident sent shock waves throughout Africa and the world which still reverberate today.
Almost immediately, before any investigation of the scene was possible, accusations were forthcoming from many uninformed sources to the effect that the then South African government must have shot down the aircraft.
For the past 28 years many lay theories have kept up the hunt for blame, without solid evidence being found to support a plot.
A thorough re-evaluation of the undisputed record reveals that the cause of the accident as postulated by many alleging foul play, is almost wholly impossible and that the loss of one of post-colonial Africa's most revered leaders was a tragic, yet preventable, accident.
The early 1980s saw the cold war between the West and the Eastern Bloc states allied to the Soviet Union at its peak.
South Africa, too, was a tense country and - externally - she was at a diplomatic impasse.
To the north of the Republic lay a belt of independent countries forming the Front Line coalition. These states were unified in their opposition to apartheid and mostly allied to the Eastern Bloc.
Mozambique had paid a heavy price in the ideological struggle being played out between the East and the West in the 1970s. After independence the western areas of the country had been used as a base by insurgents fighting the minority government of Rhodesia. In retaliation, the Rhodesian government had established a group called RNM (later RENAMO) ostensibly to act against ZANU and ZIPRA forces. The group, however, became directly engaged in an increasingly violent civil war with the ruling party of Mozambique, FRELIMO.
Once the government of Mr Ian Smith had been replaced by President Mugabe in 1980, the support and supply of weapons to MNR rebels from that quarter ceased. It was, however, continued by South Africa.
President Machel had come to realise, by 1983 that the instability his country was experiencing, was detrimental to long term development. President Machel knew that the MNR needed to be contained. That meant trying to get South Africa to stop supporting it.
In turn, South Africa was keen to stop Mozambique being used - as it saw matters - to plan and launch any possible acts of aggression against South Africa.
In 1984, after long negotiations, a non-aggression pact was concluded between South Africa and Mozambique. The agreement, called the Nkomati Accord, was signed in the town of Komatipoort on the border between South Africa and Mozambique on 16 March 1984 (the full text of the treaty can be read here.)
In President Machel, and through the accord, P.W. Botha's government had an opening to ensure that at least one part of their continent-wide northern border would be more secure. The economic benefits of the re-opening of the port of Maputo to South African commerce as a consequence of the accord was a further vital, beneficial aspect of the treaty.
It could be argued, therefore, that South Africa needed, in order to ensure continued stability and progress towards realising the potential of the Nkomati Accord, for President Machel to remain in office for as long as possible.
The accident scene
Within an hour of the impact, the South African Police had been informed of the accident by local villagers. They, in turn, alerted the air force who dispatched a specialist medical team from Hoedspruit Air Force Base to attend on the scene
In the interim, without training and the requisite equipment, the local police officers and district surgeon could do no more than to try and keep survivors warm, dry and as comfortable as possible lest they cause more harm than good through ill-advised or improper treatment.
The specialist medical team arrived shortly after midnight. It was immediately apparent that His Excellency President Machel had sustained a severe, fatal head wound on impact which precluded any possibility of survival. This is an established and uncontested fact agreed to by the South African, Mozambique and Russian pathologists who reported on their examinations. Rumours and allegations to the effect that he had survived the crash and was left to pass away by the South African authorities are, taken on the undisputed medical evidence, wholly without foundation.
Survivors were airlifted to hospital where, sadly, one was to succumb to injuries in the days ahead. As is protocol with aircraft accidents, the entire area had been cordoned off and nobody was allowed into the accident area lest items be moved or evidence destroyed.
Mr Jordaan, the Director of Aircraft Accident Investigation in the SA Department of Transport (DOT) was advised of the accident at 05.30 the following day. In terms of the International Civil Aviation Organisation (ICAO) Annex 13 The investigation of aircraft accidents - Mozambique as the state of registration and the USSR as the state of manufacture, were advised of the accident by 6.00 a.m. that morning and invited to send representatives to assist with the investigation.
The first of the South African team of investigators flew directly to Komatipoort, the aerodrome nearest the site and were ferried to the scene by helicopter arriving at about midday on 20th October. They were joined on 21st by a third investigator, Captain Roy Downes.
"I arrived at the accident site on 21 October." said Roy. "The wreckage trail began where the initial tree strike, by the port wing, occurred. There followed several other strike marks on smaller thorn bushes also made by the port wing. Shortly after, the port wing-tip struck the ground on the summit of a hill. We determined this from the evidence of pieces of the red navigation light lens, found at the point of ground contact, together with the wing outer panel."
A geological surveyor later accurately surveyed the accident site. From the tree damage and the wing-tip ground scar, he determined the angle at ground impact was approximately 10˚ at an elevation of 2193 feet or roughly 685m above sea level.
The fuselage had gouged a 51-metre long furrow in the hillside, before bouncing a distance of 42 metres. It was apparent that only after this bounce, did the major destruction of the airframe occur. In total the wreckage trail exceeded 800 meters.
"It was evident that the flaps and undercarriage were retracted. The angle of impact suggested the aircraft was in a relatively high speed, wings level, shallow descent at the point of impact. " continued Roy. "The damage sustained was certainly consistent with that of a high-speed controlled flight into terrain event."
In the days to follow both the Mozambique and Soviet authorities sent investigators to the scene. One of the Soviet investigators was Vitaliy Nisselovich* and he kept a diary of his participation. During the course of numerous interviews and e-mail exchanges, he shared observations from the diary.
"Right away we were briefed by security officers. They said not to speak in English with the South Africans. We had an interpreter go with us to all meetings." recalled Vitaliy. "All the investigators from South Africa I worked with were very honest and knowledgeable in aircraft matters. From Mozambique we had Paulo Muxango, Director of Civil Aviation in Mozambique and Antonio Neves, chief engineer of the local airline. Also these men from Mozambique were professional and correct. Everybody wanted the real technical facts."
"At the start we were told by security officers that South Africa must have shot down the aircraft. So the first time we got there we looked at all aircraft parts. There were no signs for shooting down. No bullet holes, no missile strike. Nothing at all. After this we sat together. Mozambique experts, South African experts and us. We make observations, we note, we discussed and agreed and then we wrote everything down about the technical matters and the chief of each team signed the notes together after every day- us, Mozambique and South Africa - and we all keep a copy."
Roy Downes confirmed this process but noted that it was time-consuming. "The Russians refused to stay overnight in South Africa and were ferried to-and-fro from Maputo each day. Often they would only arrive after midday. This was very frustrating." He added that one of the technical team from the South African Department of Transport was fluent in Russian and this permitted the South African team to quietly double-check the translations made in the field.
Cockpit Voice Recorder (CVR)/Digital Flight Data Recorder (DFDR) decoding
Despite insistence from Russia that the recorders be handed over to their team without delay, these were sealed and stored for safe keeping. This was due to the insistence of the South African authorities that a mutually acceptable neutral country should be used to decode the CVR and DFDR. After much debate, Switzerland was selected as acceptable to all parties for the decoding of the CVR unit.
The DFDR, however, could only be decoded in Russia and thus a dual-pronged journey was undertaken to deliver the recorders to the respective specialists in December 1986.
As was the case with all aspects of the investigation, all three investigating teams were present during the decoding process and the resultant documents of record were agreed to, and signed, by all three teams.
The Official Board of Inquiry
An official board of inquiry had been established in terms of the South African Civil Aviation Act. Although the Soviet Union and Mozambique were invited to send representatives to the board, the invitation was turned down.
The make up of the board was, however, certainly impressive. Members included: Sir Edward Eveleigh PC, former Lord Justice of Appeal in the UK. Colonel Frank Borman, the test pilot/astronaut of Apollo 8 fame, Geoffrey Wilkinson, the Chief Inspector of the UK Air Accidents Investigation Branch. Mr J Germishuys, former Commissioner for Civil Aviation in the RSA and Mr P van Hoven, Chairman of the South African Airlines Association. The Board was chaired by the Hon. Mr. Justice Cecil Margo, DSO, DFC, FRAeS, himself a former Bomber Command pilot. The Board was assisted by both South African and UK legal representatives.
Importantly, the final report - presented as a Public document and including a full transcript of the CVR - was unanimously approved by the Board members and placed in the public domain. The published report included all technical observations of the individual teams as well as a thorough evaluation of alternative theories that appeared in the public domain regarding the accident.
What was determined?
The Aircraft and crew
The Republic of Mozambique had, in the 1980s, two TU-134A-3 airliners for use of the state. This model had been designed and manufactured from 1969 to 1980 in the Soviet Union by the Tupolev Design Bureau.
While outwardly a TU-134 airliner, the inner furnishings of the Mozambican presidential jet, built to order in 1980 and registered as C9-CAA, were to a VIP standard.
As part of its package of aid and co-operation with Maputo, the Soviet Union had provided experienced Russian aircrew to fly the aircraft. The service records as well as the data recovered from the flight data recorders showed that the aircraft was fully serviceable and devoid of defects.
It became apparent that a sequence of non-mechanical factors had led to the accident and thus the board reconstructed the events of the day as nearly as was possible.
The Tupolev 134A-3 C9-CAA of the Mozambican Presidential flight photographed in Holland 2 years before it crashed. (Photo: © Gerard Helmer - Used with permission)
The Outbound Flight
Maputo to Mbala via Lusaka
The Russian flight crew reported for duty in Maputo at roughly 5.45 am local time (local time is used throughout this article) to prepare for a flight conveying President Machel to the Zambian town of Mbala for a summit meeting.
Without filing a flight plan, the aircraft departed Maputo at 7.15 am routing via Lusaka to Mbala. The landing in Luska was at 9.05 am and the aircraft was refueled. It took off for Mbala at 9:46 am.
A full fight plan from Lusaka to Mbala was lodged and this showed the aircraft's computed endurance on departure to be just on 5 hours of flying time.
Lusaka to Mbala
The landing at Mbala at 11.00 am exceeded the maximum landing weight of the aircraft. While not significant of itself, this was contrary to the expected care and diligence given the VIP passenger's profile and importance.
The extreme heat and humidity in Mbala in October - it is about 8 degrees south of the Equator - led the crew to operate the auxiliary power unit (APU) in order to make use of the aircraft's air conditioning system throughout the majority of the day as they waited for their VIP passengers. The APU is a mini jet turbine which uses fuel from the aircraft's tanks.
As matters stood the computed fuel endurance of 5 hours given at Lusaka had been reduced to just under 3 hours and 45 minutes by the flight to Mbala. Deducting the estimated flying time to Maputo of just over 3 hours, shows that the aircraft would have had less than 45 minutes of fuel aboard as a reserve without the APU consumption being deducted.
Any modern commercial or airline crew would immediately know that this 45 minute reserve is dangerously close to the minima prescribed for aircraft operations. These minima state that sufficient fuel should be available to permit the aircraft to land at a suitable alternative airport when reaching the end of its journey plus 30 minutes worth of flying time. Where no alternative is needed or available, then the 30 minutes of flying time becomes the minimum.
The return Journey
Mbala to Maputo
In the event, however, the crew did not re-fuel at Mbala nor did they decide to land at Lusaka on the return journey.
"My training would make me stop at Lusaka for fuel." said Vitaliy. "Not just to be safe for myself but I also have a very important man in the aircraft. What if there is weather at Maputo? What if I must land somewhere else? This decision was not good."
Take off for Maputo took place in the dark at 18:38. At that point the crew had been on duty for nearly 13 hours.
The aircraft overflew Lusaka before heading out over Zimbabwe. As the crew had not filed a flight plan for the return journey either, the aircraft's passage through Zimbabwean airspace appears not to have been recorded or monitored.
After arriving at the junction of South Africa, Mozambique and Zimbabwe the aircraft turned due south on course for Maputo.
At 20:49 communication with Maputo tower resulted in the weather at the airport - which included a 50% cloud cover at the time - and the expected time of arrival being confirmed. The crew were told to expect a straight in approach to runway 23 with no delays. The estimated landing time at Maputo was calculated as 21:15.
As the aircraft passed abeam of Beira, a decision would have had to be made to either continue to Maputo or to divert. On the facts, the flight proceeded even though the fuel situation on board would have indicated that the crew would have a critical reserve once they reached Maputo.
As that reserve was insufficient to permit a diversion back to Beira, the crew had basically painted themselves into a corner. They simply had to land at Maputo or declare an emergency and divert to either Matsapha or Hoedspruit - either of which would have been embarrassing for the crew.
The easiest - politically speaking - alternative to Maputo, within the limited range was Matsapha in Swaziland. As it so happens, the frequency for the VOR stations at Maputo (Coded VMA) and at Mtshapa (Coded VMS) differ only by 5 decimal points.
A VOR (Very High Frequency Omnidirectional Radio Range) signal is akin to a lighthouse with two lights. One is always shone through 360 degrees and can be picked-up wherever you may be in relation to it. A second, focused signal sweeps around the compass points at a set speed and you will only see this as it sweeps past your position.
The constant signal pulses when the narrow beam passes north. This permits you to calculate your position based on the phase difference between the pulse and the narrow signal's appearance at your position.
An instrument in the aircraft calculates where the aircraft is in relation to the VOR station using this phase calculation. A pilot can tune to a desired VOR station and set an instrument on the panel known as the Omni Bearing Selector (OBS) to indicate which direction the aircraft should be steered in order to reach the station.
It is important, however, to understand that a VOR station provides direction guidance information only and does not provide altitude guidance. It has no role to play in the decision to descend an aircraft at all.
Furthermore, a VOR station is not something one can cobble together in a secret workshop. An example of a relatively compact, modern one is shown herewith. The yellow figure provides human scale.
(Image from Wiki-Commons - Photo by Yaoleilei/GNU Open licence. Annotation by author.)
VOR Setting on the Tupolev TU134
The main frequency selector for these navigational aids on the Tupolev was located in the central instrument panel. The needle indicating the direction to fly is, however, on the navigator's display panel. On western aircraft the indicator for direction is on the main panel of the cockpit.
The main radio navigational aid tuning panel aboard a TU134 is in the central instrument panel. Note the small number windows on the VOR tuning dials (just below yellow placards) placed either side of the central selector panel.
(Image © Brian Grey Aviation Images - Used with permission. Annotation by the author.)
The TU134, however, had a secondary VOR selector placed next to the co-pilot's throttles. The TU134 was what was known as a "crew heavy" aircraft. In addition to the usual two flight crew members, there was also a navigator, a radio operator and an engineer.
In order to make the navigator's job easier, his instrument panel incorporated a switch permitting him to select the feed for his VOR display between the primary (VOR1) and the co-pilot's (VOR2) frequency. Most importantly to understanding this accident, it was possible for the navigator to initiate turns using the auto pilot system from his position.
Data recovered from the CVR indicated that the crew had intended flying towards the Maputo VOR (VMA) - located on the centre line of runway 23 and - most importantly - only 1.8 nautical miles (3km) from the runway threshold.
Approaching Maputo, there would have been a slight deviation to the left of track, to position the aircraft to intercept the VMA 045˚ VOR signal - known as a radial. Turning the aircraft onto a course of roughly 225 degrees at that point would have allowed the aircraft to intercept the instrument landing system beam for a straight-in approach to runway 23.
Clearly, the aircraft had not flown the course the crew had intended.
The early turn and decoy beacon allegations
A senior Aeroflot navigator used the decoded DFDR data to recreate the actual route flown for the investigators.
The plotted track clearly indicated that some 54 nm (95km) too early, a turn onto a heading of 225° - which would coincide with a VOR beam coming from 045 degrees - was made.
Whilst discussing the track data with the navigator, it occurred to Roy Downes that the point at which the turn was made, appeared to coincide with the 045° radial from the Matsapha VOR (VMS).
Vitaliy remembers the moment well. "The navigator went to the other team members and told us what the South African investigator had said about turning towards Matsahpa. Our KGB man immediately said that the South Africans must have tricked the crew to do this."
The heated exchange that followed between the two delegations led to the Russians explaining that it would have been a simple matter for the South Africans to change the crystals in the VOR transmitter at Matsapha.
Roy Downes takes up the story. "When Rennie Van Zyl (the senior South African official present) pointed out to the Russians that this would have been impossible as Swaziland was a sovereign nation and not part of South Africa they were taken aback and went into another huddle."
"After this the security officers always tried to make us agree to say there was another beacon. That is not logical to me at all." said Vitaliy.
Could a decoy beacon have made the aircraft turn?
Professional aviators have - from the first - called nonsense on the allegation that a decoy VOR beacon led the aircraft to impact the ground.
The most convincing argument against such a decoy beacon is that - in order to ensure that the crew tuned towards its signal and descended to a point the thought was the runway - it would have had to be extremely powerful and located within Mozambique and, most importantly, the actual signal at Maputo would have to be switched off to prevent confusion between the signals.
Vitaliy's diary notes that answers given to the Russian team by staff at the airport indicated that the NDB and VOR were sending out their signals normally throughout the night in question. This accords with the observations of the captain of a Mozambique Airlines (LAM) Boeing 737 C9-BAA, who was flying towards Maputo during the critical minutes in which the Presidential jet was approaching Maputo - and if the decoy beacon theory is to be believed - while it was picking up another beacon on that same frequency.
According to a witness statement from the LAM pilot in command, Captain Marques, the correct radar, NDB and VOR signals were all picked up by his aircraft more than 300km from Maputo.
These facts were also confirmed by a Canadian DFDR expert, Bernard Caiger, to whom the Mozambique authorities gave radar plots and the statement of Captain Marques for independent analysis. He checked the plot of the B737's flights and found no deviation from its route as would later be implied by the minority report generated by the Russians.
In order for a false VOR to have any hope of success in changing the course of the aircraft, it would have had to be radically more powerful than the Maputo signal and this would have made the LAM airliner deviate to the west.
Nothing of the sort took place.
Possible reasons for the early turn
Given the fuel situation which was approaching a critical point - the low fuel warning light had illuminated for 25 seconds at the top of descent a minute or two prior to the turn on to the 225 degree heading - it is entirely possible that the Captain, concerned at the low fuel situation, may have tuned to the Matsapha VOR station to determine the distance using the Distance Measuring Equipment (DME) in order to check his options.
It is possible that, while the signal was coming in from Matsapha, the navigator saw the VOR needle indicating a turn to the right. He did not, however, have the frequencies displayed at his position - only the direction indication. In the event this poor layout of the flight deck undoubtedly played a major role in the accident.
When the turn was made without his command - and having just been given a distance of some 100km still to run to Maputo - the Captain was surprised.
21:11:28 Captain: Making some turns? Couldn't it be straight?
21:11:32 Navigator: VOR indicates that way.
It is significant that no further discussion of this early turn ensued on the flight deck, Logic and cross-checking by the crew members would have alerted them to the fact that, with at least 100km still to fly to Maputo, in addition to being at a much higher altitude than usual when the turn was normally made, something was amiss.
Another possibility is provided by Roy Downes. "The number one VHF/VOR/DME frequency selectors and Omni Bearing Selectors (OBS) located on the central instrument panel were WWII vintage, or earlier. They consisted of an engraved metal disc, turned within a rotating housing."
Setting these selectors was it seems, not a simple process.
"The only lighting for the selectors was by means of the cockpit floodlights. The back lighting the western world accepts as normal was simply not present. A most disturbing feature of the engraved decimal figures to my eyes was the similarity between the figures 3 and 7. The top line of both figures was a straight line, while the lower curves of the figure three were not accentuated to the same degree, as is normal with the western figure three. Easily confused."
With antiquated equipment such as this, especially in conditions of poor lighting, he believes an inattentive crew member could easily err between the two figures.
It is possible that the co-pilot or Captain could have inadvertently set the wrong frequency and later, while they were checking the navigation aids due to their confusion at not having reached Maputo, reset it to the correct one.
The Maputo VOR frequency was 112.7 MHz while the Matsapha VOR frequency was 112.3 MHz.
"For this reason alone, the identifying of all selected radio navigation aids, by means of the Morse code identifier is essential. The CVR clearly indicated that this crew did not bother with this important detail which, according to their Standard Operating Procedures was clearly a requirement." Roy pointed out.
The CVR recorded that there were numerous things the flight crew omitted. The top of descent checklist was one. Instead, a discussion on the flight deck in which arrangements were made with the cabin crew to provide the crew with some soft drinks and imported beers from the bar for later use occupied their time. The co-pilot was listening to the news on the radio. The Captain was looking for a pen.
Had the crew been drinking?
Post mortem tests on the pilots did not indicate any alcohol in their systems at all. Many media reports were published in South Africa shortly after the accident to the effect that he crew were drinking in the cockpit but a study of the factual evidence shows this to be false.
The discussion of the drinks, however, used up valuable time at a stage of the flight when the crew should have been focused on the approach and in ensuring that they had completed their pre-landing duties.
The final minutes
As has been said, the decision to turn an aircraft in response to a VOR indication is not automatic unless the auto pilot system has been told to lock on to a specific VOR station and fly automatically towards it.
Descent of an aircraft can only be automatically controlled if the aircraft is fitted with an instrument landing system (ILS). This system will then only issue such commands if it is engaged and the aircraft has made its final turn towards the runway - normally just a few kilometers from the airport. Any decision to descend when still far away from an airfield - such as was the case with C9-CAA - can only be made by the crew. In all cases the rate of descent and altitude should always be monitored by the crew.
Of supreme importance to the final minutes is the lack of cohesion as regards procedures and resource management of the crew.
At no stage were any checklists read-out or performed. After the turn the crew should have begun its landing briefing and performed the appropriate checklists given that they were convinced they were approaching to land. This was not done despite their impression that they were very close to Maputo.
Multiple clues should have been followed-up by the crew to cross-check their situation. One or two simple questions to the tower at Maputo would have alerted them to the fact that their impression of their position was incorrect. As can be seen from the CVR record, the disaster could have been averted at a number of key points.
21:09:16 the Navigator tells the Captain they have 120 km to go.
21:10:48 the Navigator tells the Captain there is 100 km left. (This meant, on his information, it took them 32 seconds to cover 20 km)
In the following minutes the Captain is searching for a pen and the crew are dividing up the beers and soft drinks they had by then obtained from the bar.
21:12:51 The co-pilot notices that the VOR indicators are not correctly lit (they were indicating the VOR station was to their left) and comments on this. Instead of getting an answer or any of the other crew responding, the navigator calls out 80km. Despite no change having been made to the descent and airspeed, the second 20km leg had taken more than a minute and a half longer than the previous 20km. This was an indication that things were going awry and was caused by the angle at which the aircraft was now flying relative to the origin of the distance signal.
If they were turning towards Maputo as they assumed, the indicator lights would not have been displaying that the chosen VOR station was to the left nor would there have been such a huge difference in the time taken for the previous 20km to be flown and the much longer time for the next 20 km. All these items should have raised warnings. Nobody picked up on this at all.
The co-pilot repeated his concern about the VOR panel lights. "They should be lit is that not so?"
21:13:05 The Captain asks "Do we always have it like that?"
At this stage, a fully focused crew would have re-assessed their information but this crew continued to discuss the drinks situation.
21:14:57 The Navigator gives a distance of 60km to run which upsets the Captain who had previously thought they would land a minute early and impress their VIP passenger.
The low fuel warning light had, by now, come on permanently. This must have increased the tension for the Captain and crew.
At this stage the Captain should have put the early turn, the VOR indicator lights indicating that their station was to their left plus the apparent changing ground speeds together. If he had added that to the fact that they had not reduced speed at all and were still running at descent speed he should have quickly discovered that something was amiss. He merely re-estimated his arrival time to be in 5 minutes time as the aircraft continued descending.
At 21:17:21 The Captain made a statement that should have provided a resounding clue that there was a problem. The Captain, looking ahead, could not see the runway he expected to be dead ahead.
"There is no Maputo" he stated. The co-pilot asks "What?"
At 21:17:27, the Captain decides that he knows the answer to their problem, a power failure at Maputo. This was a totally incorrect assumption given the Tower was able to communicate with his aircraft and the airport had standby power for the radios and lights in any event. Nevertheless he states flatly "Electrical power is off chaps."
21:17:42 At this point the navigator had begun to realise things were not adding up and he had also noted that the ILS and other items were not picking up on his instruments as he expected. This was due to the actual position of the aircraft, some 60km away from the airport. "There is something that I don't understand, Ahh..."
21:17:49 The navigator adds "ILS switched off and DME."
The aircraft continued descending but, given that their estimated arrival over the runway threshold had not taken place yet, the Captain was becoming aware something was amiss and instructed the Radio officer to call the tower.
19:18:24 The aircraft passed 3000 feet and the radio operator called the tower to report "Maintaining 3000 feet."
This was the minimum safe altitude over Maputo and should have been maintained until the airfield was in sight. The aircraft was, however, not maintaining its altitude. Nobody picked up on this or did anything about it. Still in descent mode, the aircraft was approaching the ground rapidly.
In the ensuing minutes, the final set of circumstances that meshed with all the others led to numerous misunderstandings between the control tower and the aircraft. When the aircraft made contact, the tower controller assumed they had the airfield in sight, but this was not the case. The aircraft was cleared to land even though it had not reported itself on final approach yet.
At 21:19:32 the navigator gave an update on the distance which alarmed the Captain.
The navigator told the Captain "25-30km". The Captain queried this and then, perhaps realising that the distances and elapsed times did not add-up with the situation some 6 minutes beforehand, stated "Something is wrong chaps."
The aircraft was left to continue descending.
Any fully alert pilot would, at that point, have throttled up and climbed the aircraft away until he had his bearings and was fully aware of his position. The captain and crew had, by that point, been on duty for roughly 16 hours. Fatigue must have been a factor.
Still persisting with the belief there was no power at Maputo (despite the radio contact being maintained with the tower), the Captain asked that the tower controller check the operation of the runway lights.
The radio call was "Maputo, check your runway lights." and not a more unambigious "Are the runway lights at Maputo switched on?"
"Check the runway lights." was taken by the controller as the crew seeing (i.e. checking") the runway lights and he cleared them to land again.
21:19:40 The Captain states again "There is a problem".
The aircraft continued to descend.
21:20:22 The Captain again thinks out loud. "I understand nothing!"
21:20:23 The Radio operator asks "Don't you see the runway yet?"
21:20:28 Captain "What runway. What are you talking about?"
21:20:32 Navigator "We are going to do straight in approach."
21:20:35 Captain "We are doing straight in approach."
21:20:38 Navigator "No well, can you see the runway?"
The co-pilot and Captain, unable to see any runway, nor indeed the city itself, came to the conclusion that it was cloudy. The standard procedure for cloudy conditions would have been to climb away and re-execute the approach and not try and maintain a visual landing process. During the period after climbing away all the navigation aids, instrument landing systems and the radar set could have been used to confirm the position of the aircraft.
None of this was done and the aircraft was left in descent configuration.
During the playback of the CVR in Switzerland, all three teams from the field investigation were present. "At the 21:20.55 and while the radio operator was making a transmission to Maputo, an alarm sounded." explained Roy Downes.
"After sounding for some seventeen seconds the captain uttered a mild expletive, apparently in response to the alarm, but took no action. To the Russians, this sequence gave rise to considerable concern and one or two, gasped in apparent amazement. The alarm continued to sound for a further fifteen seconds before it ceased abruptly. Seven seconds later, the aircraft collided with the rising terrain."
"This sound was ground proximity warning system" explained Vitaliy "The Captain hears this but carries on letting the aircraft fly downwards. No throttle action was taken. No climbing orders. No discussion with the crew. This was the most surprising point of the entire accident for me. It also assures me that nobody outside the aircraft made it crash."
When pressed about this view, Vitaliy pointed out that, in order to succeed with the alleged plot to lure the aircraft off course and kill the President of Mozambique, aside from all the previously discussed technical factors that needed to line up, the conspirators would have needed to ensure that the 5 crew members would dispense with normal flying discipline as well as ignore the many warning signs that could have alerted them to their pending peril - particularly the final one.
"Make him ignore ground warning system? For this you must make the Captain want to die, the co-pilot and the others. You must let the others all agree too. You must make them take no action to save everyone when they had the time to save the flight. This kind of mind control I have never seen. Such a thing is not possible."
Roy Downes summed up the matter thus: "There can be little doubt that this crew was, during the final minutes of the flight, utterly confused. They had made a number of cumulative errors, got off course and eventually realised that something was amiss. Yet, instead of maintaining a safe altitude, they blundered on down, ignoring a final warning to their peril, to become another Controlled Flight Into Terrain (CFIT) statistic."
Given all the non-aviation voices that have commented on the accident and which have clouded the true facts regarding this accident- despite the factual aspects being totally aligned between the three nations that investigated the matter - anyone keen on fully understanding the accident is encouraged to read the full report that is available on the CAA website prior to repeating the unfounded, and unsupported, allegations of foul play.
This map provides a bird's eye view of the sequence of events involving the loss of C9-CAA. From this perspective the commonalities between the path the crew believed they were following and the actual one is easily apparent. It is also easy to see how the paths of the VOR radials of Matsapha and Maputo mirror each other. The black line between the two flight paths indicates the estimated time of arrival after which the crew should have become acutely aware that something was amiss. Sadly, they continued descending below the minimum safe altitude and impacted the high ground near Mbuzini. (Map by the author)
* Given the sensitivities of the matter and the official Russian position as evidenced in their minority report at the time, the investigator from the former Soviet Union requested that a nomme-de-plume be used. The author has respected this wish.
The author wishes to convey his thanks to the CAA, the Mozambique aviation community, the community of Airliners.net, Roy Downes, Brian Grey and the many other aviation professionals contacted during research for their provision of information, contacts, images and other material that made this article possible.
Flight Safety author Mark D Young has been writing on aviation safety for more than 23 years. He has published a book on the history of SAA accidents called "A Firm Resolve" and is currently working on a further book regarding notable international aviation accidents. For more details on his work see www.flightlevel42.co.za/flightsafety.htm
Click here to sign up to receive our free daily headline email newsletter