Mid-Air Collision over Yugoslavia 1976

Forum for discussing non-Pakistani aviation issues and news.
smhusain_1
Registered Member
Posts: 225
Joined: Sat Jul 19, 2008 1:16 pm
Location: Brampton, Ontario, Canada

Mid-Air Collision over Yugoslavia 1976

Post by smhusain_1 »

“The town of Split lies on the coast of Yugoslavia almost exactly between Italy to the north and Albania to the south. Since earliest times travellers have been attracted to its sunny shores and the warm Adriatic Sea. In the third century AD the Roman Emperor Diocles built a magnificent fortified palace for his retirement on an attractive cove which now forms the harbour of the modern town. Many parts of the building can be seen to this day. In the Emperor’s footsteps have followed armies of tourists from the cold north, seeking sun and relaxation in the pleasant resort. West German holiday makers, especially, have found it a popular destination.

On the morning of 10 September 1976, a party from Cologne holidaying in the Split area prepared for their return home after an enjoyable stay. As they faced the trauma of check-in at busy Split Airport, another group of travellers was already beginning its journey from the distant airport of London, Heathrow. British Airways scheduled flight BE476 from London to Istanbul lifted off from Heathrow at 08.32 hrs GMT (09.32hrs local). The Trident 3B, registration G-AWZT (Zulu Tango), was less than half full for the 3 1/2 hour journey to Turkey, with only 54 passengers on board. The travellers formed a disparate group of various nationalities and even included one stateless person. Captain Dennis Tann commanded the Trident with First Officer (F/O) Brian Helm as co-pilot and F/O Martin Flint in the P3 position. Six cabin crew tended the passengers’ needs. Flight BE 476 crossed the channel at Dover, then proceeded south-east over Belgium, passing overhead Brussels, and continued to West Germany. About 1 ¼ hour after take-off, with breakfast completed, Flight BE 476 passed over the city of Munich and reported its position at 09.43 hrs. The Trident settled on airway Upper Blue 1 (UB1), flying at flight level 330 (thirty three thousand feet above mean sea level) towards the radio beacon of Villach in the south of Austria. At such quiet moments in the cruise, with eyes and ears still alert, crews often keep their minds active with the daily cross-word puzzle. F/O Helm was still struggling with a clue and engaged the assistance of the others. As the quiet discussion ensued Trident Zulu Tango slipped from West Germany into Austrian airspace at 09.48 hrs. At precisely the same moment, Flight JP 550, a DC-9 of the Yugoslav charter airline Index-Adria Aviopromet, lifted off from Split airport carrying the party of tourists returning to Cologne. The DC-9, registration YU-AJR (Juliet Romeo), was almost full with 108 passengers, of whom all, except one Yugoslav, were West Germans. On the DC-9’s flight deck were two pilots, Captain Joze Krumpak and F/O Dusan Ivanus. In the cabin the passengers were attended by only three stewardesses who had their work cut out catering for so many people on the short 2 ½ hour journey.

Flight JP550 on its northwesterly route towards Cologne and Flight BE476 on its southeasterly course towards Istanbul, were both flight-planned to traverse the busy area over the Zagreb radio beacon. A glance at a map of the region clearly indicates Yugoslavia’s important position in aviation, as air traffic between Europe and the East routes south of the Eastern Bloc nations. Charter flights from Northern Europe in summer to Yugoslavia, Greece and Turkey also swell the flow of traffic. Five high level airways criss-cross the Zagreb region like a disjointed Union flag with three airways—Upper Blue 5 (UB5), Upper Blue 9 (UB9) and Upper Red 22 (UR22)—intersecting over the Zagreb radio beacon. A fourth airway, Upper Blue 1 (UB1), routes to the south of Zagreb while Upper Amber 40 (UA40), the fifth airway, originates at Zagreb on a direct to Sarajevo. In the previous five years a rapid and sustained growth of air traffic over Western Yugoslavia had resulted in over 76,000 aircraft movements being handled by Zagreb Air Traffic Control Centre (ATCC). By 1976 the Zagreb was the second busiest in Europe and 30 controllers were struggling to handle traffic which required more than double that number. An extensive training programme had been introduced but the centre was still desperately short of experienced controllers. Three years earlier a modern radar flight control system had been installed but it had not yet been properly calibrated and was considered unreliable. It was claimed that the Swedish equipment did not meet contract requirements. As a result the radar system had not been properly commissioned and was not depended upon for aircraft separation. The Zagreb ATCC relied very much on procedural control with pilots transmitting their positions at specified reporting points along the airways. These position reports were then monitored by the radar system.

The increased work load for the Zagreb staff in the years of air traffic growth produced a very difficult working environment. Lapses in concentration were perhaps not surprising and a number of incidents had resulted, with 32 near-misses being experienced over the last five years. Two controllers had been dismissed for negligence. On a lesser scale lapses in discipline resulted in lateness for duty, and unauthorized absence from the control station. These failings, however, had to be viewed in the light of trying circumstances and in the very high volume of traffic which had passed safely through the region.

On the morning of 10 September 1976, as BE 476 and JP 550 converged on the Zagreb area, the staff at the Zagreb ATCC faced another day of heavy traffic. The airspace over Zagreb was divided into three distinctly separate layers—lower, middle and upper—with middle and upper sections each directed in the short-staffed centre by a controller and his assistant, where normally three personnel—radar controller, procedural controller and assistant controller—would be required. JP 550, on its climb over the Zagreb radio beacon to its planned flight level of 310, would pass through the middle control layer (from 25,000 to 31,000 ft), while BE 476, already cruising at level 330, would pass through the upper control section (above 31,000 ft). On the duty shift that morning, under the supervision of Julije Dajcic, was five controllers responsible for the middle and upper stratum of airspace. Apart from the 43-year old Dajcic, their ages ranged from late 20s to early 30s. The controllers worked a 12-hour duty day with normally 2 hrs at a control station followed by a one hour break. By 10.00 hrs GMT (11.00 hrs local) the morning shift, which started at 0700 hrs local, had already been on duty for four hours. At the middle section console sat Bojan Erjavec as controller, who had been at his station for the past hour, with assistant controller Gradimir Pelin, who had just started duty. Mladen Hochberger controlled the upper section and was due, at that moment, to be relieved by Nenad Tepes. Hochberger’s assistant for the past hour had been Gradimir Tasic who had already spent the first two hours of the shift as duty controller. When Hochberger’s relief, Tepes, who was now late, arrived Tasic would also act as his assistant for the next hour, monitoring procedures and coordinating flights with other regions on ground telephone links. Of the staff mentioned, all except Tasic, had had a period of at least 24 hours off duty in the last few days. Tasic was on his third day in a row of 12 hours on duty.

As the Inex-Adria flight climbed out of Split, the approach controller who, in spite of his title, also handled airport departures, was having trouble co-ordinating JP550’s ascent with Zagreb lower east sector. As a result the Yugoslav DC-9 was required to cross the Split radio beacon at level 120 before continuing its climb to level 190 while proceeding towards the radio beacon at Kostajnica. At 09.55 hrs GMT JP550 established contact with the Zagreb lower controller and was further cleared to level 240. One minute later, the clearance to climb to level 250, in the middle level airspace, was received by JP550. A request for the DC-9 to call passing level 220 was also made as a reminder for the lower controller to instruct JP550 to change frequency to the middle controller at that point. JP550 had flight planned for level 310 but levels higher than 260 in the middle sector were not available because both levels above were blocked by other traffic. Westbound jet cruising flight levels are 260, 280, 310, 350 and 390 while eastbound levels are 270, 290, 330, 370 and 410. In the middle sector, Adria 584 from Split to Nuremberg was at level 280, estimating Zagreb at 10.08 hrs and Olympic 187 from Athens to Vienna at 310, estimating overhead Zagreb at 10.11hrs. At that time JP550 was estimating Zagreb at about 10.16hrs. At 10.02hrs JP550 radioed passing level 220 and was instructed to transmit the next call on the middle sector’s frequency at 135.8 MHz. After the frequency change the DC-9 crew waited about 30 secs for a break in the flow of conversation then established contact with Erjavec, the middle sector controller.

10.03:21 JP550 R/T: ‘Dobar dan [Good day], Adria 550, crossing two two five, climbing two six zero.’
10.03:28 Zagreb middle (Erjavec) R/T: ‘550, good morning, squawk alpha two five zero six, continue climb two six zero.’
10.03:34 JP550 R/T: ‘Squawk alpha two five zero six, continue climb two six zero.’
10.03:38 Erjavec R/T: ‘That is correct. Inbound Kostajnica, Zagreb, Graz next.’

Alpha 2506 is one of the radar identification codes used in secondary radar operation. With primary radar a signal pulse is simply transmitted from a ground station and reflected back from a target to the station as a weak echo. Secondary radar, on the other hand, involves a ground transmitted radar signal being received on board an aircraft by a small receiver/transmitter known as a transponder, which responds by transmitting a second signal which is in turn received by the ground based radar receiver. The resultant signal on the ground is much stronger than the weak echo detected using primary radar. An added advantage is that the transponder transmits on a frequency different from the ground based radar transmitter. Since the radar receiver is tuned to the transponder frequency, weak echoes of the transmitted radar signal that may be reflected from the target or from storm clouds in the vicinity are eliminated on the radar screen. A distinct clear image of the target is displayed. On instruction from the control the transponder is programmed by the pilots selecting a four number code, and the procedure is known as squawking. Each aircraft is allocated its own code which displays on the radar screen by the target a flight label consisting of the squawk code, flight number, and the aircraft flight level. As a further refinement the centre at Zagreb used a radar height filtering procedure where only certain codes were assigned to each control layer. In any airspace layer, only aircraft with appropriate codes (middle sector 2500-2577, upper sector 2300-2377) showed up with a flight label on that controller’s radar console. All other aircraft in the vicinity, higher or lower, showed up only as target blips. As a precaution, however, if an aircraft strayed unannounced in another layer, for example from the middle to the upper section, then the flight label would in that case appear automatically when the aircraft climbed above level 315. A controller could also obtain the flight label on his screen of an aircraft outside his airspace layer in order positively to identify an unlabelled target. This could be achieved in a number of ways, but the simplest method was to position a ‘pointer’ on the radar screen over the unidentified blip and then request the computer for information. The identification label would then appear by the selected flight for 30 sec. This procedure, however, took time, and involved the use of a separate key board thus distracting the controller from radar monitoring.

The hand-over of JP 550 from the lower east sector to the middle sector during its climb was completed in good time, which was just as well. Contrary to instructions no progress strip with JP550’s flight details had been prepared in advance for the middle sector. The flight progress strip is about eight inches by one inch and contains important information on a particular flight including call sign, aircraft type, requested flight level, airway routing and squawks code. The strips are mounted on metal backing plates which are slotted in racks by the controller. Checking, altering, and arranging the strips are part of the assistant controllers’ tasks. The Zagreb middle controller had to adjust quickly to accept JP550 into his sector without prior knowledge of the flight, but since the changeover was conducted in the lower airspace, the transition was accomplished with safety.
BE476 crossed over the Klagenfurt radio beacon on the Austrian/Yugoslav border at 10.02 hrs, about the time of JP550’s exchange, and was instructed by Vienna Control to contact Zagreb on frequency 134.45 MHz. As the Trident flew eastbound on airway UB5 towards Zagreb at level 330, all was not well at the Zagreb upper control station. Controller Hochberger, still waiting on his relief, was showing his impatience with his colleague’s late arrival and vacated his seat to go and look for Tepes. By the time BE476 called Zagreb upper, Tasic, who was supposed to be in the position of assistant, was now completely on his own, controlling flights, monitoring procedures, and liaising with adjacent control centres. Although at 28 Tasic was the youngest in the room he was an experienced and competent controller, but the flow of air traffic through his sector was as much as any man could handle.

10.04:12 BE476 R/T: ‘Zagreb, Bealine 476, good morning.’
Zagreb upper (Tasic) R/T: ‘Bealine 476, good morning, go ahead.’
10.04:19 BEA 476 R/T: ‘476 is Klagenfurt at zero two, flight level three three zero, and estimating Zagreb at one four.’
Tasic R/T: ‘Bealine 476, roger, call me passing Zagreb, flight level three three zero, squawk alpha two three one two
10.04:40 BEA 476 R/T: ‘Two three one two is coming.’

Previous radar stations had recorded the Trident level precisely 330, but on Tasic’s screen the flight label indicated BE476 to be at 332 (33,200 ft) or 335 (33,500 ft). Since the equipment was not considered to be completely reliable the discrepancy was ignored.

10.04:11 TK 889 R/T: ‘Zagreb, Turkair 889, over Charlie, three five zero.’
Tasic R/T: ‘Turkair 889 contact Vienna control one three one er . . . sorry, one two nine decimal two (129.2 MHz). Good day.’
10.04:54 TK889: ‘one two nine decimal two. Good day, sir.’

On the Trident flight deck the crew watched the opposite direction Turkish aircraft flash past above them at level 350. ‘There he is’, called one of the crew.

Tasic now required onward-clearance from Belgrade for an Olympic Airways’ flight proceeding eastbound on UB1 towards Sarajevo. He called the Belgrade ATCC on the ground telephone link to speak to the relevant upper controller’s assistant.
Tasic Tel: ‘I need Sarajevo upper.’
Belgrade Tel: ‘Right away?’
10.05:17 OA 182 R/T: ‘Zagreb, Olympic 182, passing Kostajnica at zero five, three three zero, estimate Sarajevo at one seven.’ (Olympic 182 eastbound on UB1)
Tasic Tel: ‘You can hear the message over the phone.’
10.05:20 Tasic R/T: ‘Olympic 182 contact. Olympic 182 report passing Sarajevo.’
10.05:25: Belgrade Tel: ‘Hello?’
10.05:28 Tasic Tel: ‘Hello, hello, listen, give me the controller . . .’

No sooner had the controller taken the phone that another aircraft called.
10.05:30 9KACX R/T: ‘Zagreb, Grummen 9KACX with you, flight level four one zero.’

Tasic ignored the call and continued with the telephone conversation.
10.05:35 Tasic Tel: ‘Er, Lufthansa 360 and Olympic 182 – they’ve got nine minutes between them. Is that OK for you?’

Normally along route should be ten minutes apart but Tasic hoped the Sarajevo upper controller would accept them with the reduced separation.
Belgrade Tel: ‘I’ve got it . . . OK.’

The incoming radio calls continued.

10.05:44 IR777 R/T: ‘Zagreb, this is Iran Air triple seven, good . . . morning.’
10.06:15 OM148 R/T: ‘Zagreb, Monarch 148, we checked Kostajnica zero five, level three seven zero, Sarajevo one nine.’ (Monarch 148 on UB1 eastbound)
10.06:37 9KACX R/T: ‘Zagreb, Grummen 9KACX is with you, level four one zero.’

Meanwhile, in the middle sector, JP550 was just levelling at 260. His position was 34nm south of the Kostajnica radio beacon, heading north due north on UB9, and estimating overhead Kostajnica at just after 10.09.

10.05:57: JP550 R/T: ‘Adria 550, levelling two six zero, standing by for higher.’
10.06:03 Zagreb middle (Erjavec R/T): ‘550, sorry three three zero . . . eh . . . three one zero is not available, two eight zero also. Are you able to climb maybe to three five zero?’
10.06:11 JP 550 R/T: ‘Affirmative, affirmative. With pleasure.’
10.06:13 Erjavec R/T: ‘Roger, call you back.’
10.06:14 J P550 R/T: ‘Yes sir.’

On leaving the control room in search for Tepes, Hochberger had met his replacement on the way and the two men stopped to discuss t5he air traffic situation. Contrary to instruction they continued to complete the hand-over outside the control room. Tasic, still at the upper control station on his own, was just about managing to handle the volume although the pressure was beginning to tell. JP550 remained level at 260 while Erjavec attempted to attract Tasic’s attention for clearance to climb the Yugoslav DC-9 to level 350 in the upper sector. Climbing aircraft through the layers of airway traffic travelling in both directions requires careful coordination by all concerned. Erjavec was satisfied with the situation in the middle sector, but climbing JP550 up through the opposite direction traffic at 330 and checking separation with other aircraft already cruising at 350 was Tasic’s responsibility. Erjavec’s raised hands were seen by Tasic who, in an obvious gesture, waved him away. All his attention was needed for his own traffic. The middle and upper control stations were only a half a metre apart so Erjavec’s assistant, Pelin, who held a radar licence, simply moved across to Tasic’s screen. Pelin pointed at JP550’s unlabelled target and asked Tasic for clearance. The confused exchange was only a brief moment in the sequence of events but proved to have far reaching consequences for those concerned. As far as Tasic understood the situation, he was only being indicated an aircraft in the vicinity of Kostajnica, whereas Pelin assumed climb clearance had been received. He moved back to Erjavec with the all clear.

10.07: 40 Erjavec R/T: ‘Adria 550 recleared flight level three five zero.’

At the upper control station Tasic struggled on alone.
10.07:45 Tasic R/T: ‘Beatours 778, squawk alpha two three zero four.’
10.07:50 BE778: ‘Alpha two three zero four coming down, 778.’

Again he was on the line to Belgrade for BE778’s clearance, but interrupted the call to reply on the radio.

10.08:26 Tasic R/T: ‘778 radar contact, continue.’

Belgrade was having trouble retrieving Beatours information. Eventually the details were found and BE778’s routing from London to Istanbul confirmed. ‘How do you spell Constantinople?’ asked Belgrade, using the old name. This was too much for Tasic and he replied in exasperation with the phonetic spelling of that city’s international four-letter designator. ‘Lima Tango Bravo Alpha.’

There was still no sign of Tepes taking up the controller’s position. In the middle sector Erjavec continued to monitor JP550’s climb and confirmed by radar that the DC-9 was approaching Kostajnica. Captain Krumpak was instructed to proceed via Zagreb and Graz and to call passing level 290.

10.09:49 JP550’s R/T: ‘Zagreb, Adria 550 is out of two nine zero.’
10.09:53 Erjavec R/T: ‘Roger, call me passing three one zero now.’

As the Yugoslav DC-9 ascended to 350, the BEA Trident 3B, BE476, cruised at flight level 330 with a true airspeed of 480kt. A southwesterly wind aloft at 45kt resulted in a groundspeed of 489kt. The estimate for overhead Zagreb remained at 10.14 hrs. In spite of the wind direction the Trident had tracked a little to the south of the airway and at 10.11:41 turned left from a heading of 121 degrees on to a heading of 115 degrees to home directly over the Zagreb radio beacon. JP550’s rate of climb was 1,800 ft/min and the 2,000 ft ascent from level 290 to level 310 was achieved in a period of about 1 ¼ minutes. The DC-9 passed into the upper control sector.

10.12:03 JP550 R/T: ‘Zagreb, Adria 550, out of three one zero.’
10.12:06 Erjavec R/T: ‘550 for further Zagreb one three four decimal four five (134.45 MHz). Squawk stand-by and good day, sir.’
10.12:12 JP550 R/T: ‘Squawk stand-by, one three four four five. Good day.

Erjavec had instructed JP550 to squawk stand-by on the transponder in order to release the DC-9’s middle sector squawk code for another aircraft. Although a not acceptable procedure, the selection of stand-by laid the foundation of a lethal trap for the unsuspecting Tasic. The radar computer was programmed to activate automatically on the upper controller’s screen the flight label of any aircraft transmitting a middle sector transponder code when it climbed above level 315.With the squawk code in stand-by mode no flight label would automatically appear. Tasic’s first task on being contacted by JP550 would be to allocate an upper sector transponder code. On the DC-9 flight deck the frequency was changed, and once again the crew waited for a gap in the flow of radio conversation before attempting to contact Tasic.

10.11:53 Tasic R/T: ‘Finnair 1673, go ahead now, copy 1673, go ahead.’
F1673 R/T: ‘Finnair 1673 passed Graz at one zero, level three nine zero, estimate . . .’ (Southbound from Austria)
10.12:10 Tasic R/T: ‘Finnair 1673, report passing Delta Oscar Lima. Maintain level three nine zero, squawk alpha two three one zero.’
10.12:20 F1673 R/T: ‘Will report passing Dolsko.’

A few seconds before the last Finnair transmission, Tasic received from Pelin JP550’s flight progress strip which once again, contrary to regulations, was simply an alteration of the one already held by the middle sector. A new strip should have been prepared by Peeling. Tasic had hardly time to glance at the information. The upper sector station had not received a strip in advance because JP550 was originally flight planned for level 310 which lay in the middle sector. It was not known until the last minute that the DC-9 was sufficiently light to climb to level 350. Unannounced JP550 penetrated the upper sector airspace and showed only on Tasic’s screen as an unlabelled target.

10.12:24 LH 310 R/T: ‘Lufthansa 310, Sarajevo at zero nine, three three zero, Kumunovo three one.’
Tasic R/T: ‘Lufthansa 310, contact Beograd one three four decimal four five (134.45 MHz) – sorry – sorry, one three three decimal four five (133.45 MHz). Good day.’
LH 310R/T:’Good day.’
10.12:38 Tasic R/T: ‘Good day.’

Just after Tasic received JP550’s flight progress strip Tepes eventually arrived. He had completed the hand-over from Hochberger outside the control room but would need to be updated by Tasic on the latest situation. Tepes was rostered as the controller for this duty period so with Tasic at the controller’s station, they would have to change seats. As the two waited for an appropriate moment Tasic, controlling aircraft in his sector, briefing Tepes in the gaps between, and still coordinating flights on the telephone with Belgrade, was reaching saturation point. There were eleven aircraft in the upper sector. Having received JP550’s flight progress strip at such a late stage, Tasic had no time to examine the details or to appreciate the developing situation. A further error confused the situation. Marked on the strip was JP550’s level of 350 but omitted was an arrow pointing upwards beside the figures to indicate the aircraft still climbing. As JP550 continued on its ascent to level 350, heading 353 degrees on UB9 towards Zagreb, Tasic was unaware of its height. The DC-9 had still to pass through the opposite direction flight level of 330. The southwesterly wind at that level increased JP 550’s true airspeed of 440 knots to a ground speed of 470 knots. At that rate, it would cross Zagreb still climbing, at 10.14 hrs, at the same time as BE 476.

10.12:40 OA 172 R/T:’Zagreb, Olympic 172, good morning, level three three zero.’
10.12:48 Tasic R/T: ‘Olympic 172, go ahead.’
OA 172 R/T: ‘Olympic 172, level three three zero estimate Dolsko at one six.’
10.13:00 Tasic R/T: ‘Olympic 172, report passing Dolsko, flight level three three zero. Squawk alpha two three zero three.’
10.13:07 OA172 R/T: ‘Olympic 172, squawk two three zero three. Report passing Dolsko level three three zero, and after Dolsko direct Kostajnica?’
10.13:15 Tasic R/T: ‘Affirm, sir.’
10.13:18 OA172 R/T: ‘Thank you.’
10.13:19 BE932 R/T: ‘Zagreb, Beatours 932 is level three seven zero, estimate Dolsko one eight.’

Tasic was distracted again with the attempt to change places with Tepes who was now filling the post of assistant and was in contact with Belgrade on the ground telephone link. Beatours repeated the call and Tasic misheard.
10.13:34: BE932 R/T: ‘Zagreb, Beatours 932.’
Tasic R/T: ‘962 go ahead.’
10.13:42 BE 932 R/T: ‘Beatours, three seven zero, estimate Dolsko at one eight.’

Once more Tasic’s attention was drawn to other matters.

10.213:53 Tasic R/T: ‘Beatours, maintain level three seven zero and report overhead Dolsko. Squawk alpha two three three two.’
BE932 R/T: ‘Roger, two three three two. (A2332)

Still no message had been received from the DC-9 in spite of the short breaks in radio transmissions. The Yugoslav aircraft had 2,500 feet left to go to its cruising level of 350 but was only 500 feet below the opposite direction Trident at level 330.

The two aircraft were both almost overhead the Zagreb radio beacon and were dangerously close. At last the DC-9 made initial contact with the upper controller.

10.14:04 JP550 R/T: ‘Dobar dan, Zagreb, Adria 550.’
10.14:07 Tasic R/T: ‘Adria 550, Zagreb, Dobar dan. Go ahead.’
10.14:10 JP550 R/T: ‘Three two five crossing, Zagreb at one four.’

Tasic could hardly believe his ears. Did JP550 say crossing level 325? Anxiously he called back.

10.14:14 Tasic R/T: ‘What is your present level?’
10.14:17 JP550 R/T: ‘Three two seven.’

A possible collision between the Trident and the DC-9 was only seconds away. Neither flight crew was aware of the impending danger but both would be maintaining a normal watch, although spotting another shining aircraft in that bright morning sky would have been very difficult. The Trident was leaving a long contrail which the DC-9 crew may have seen, but judging another’s height especially with the aircraft nose up in a climb is almost impossible. Also the sun was behind the Yugoslav aircraft and eye contact for the Trident crew would have been hard to establish. The ability of the eye to detect other traffic is improved when two aircraft move relative to each other. The important factor of an impending collision is that a line of constant bearing between two aircraft involved in an encounter and the phenomenon makes it more difficult for them to spot each other. On the Trident flight deck JP550’s radio message would have alerted the crew to the danger, but since position reports are normally passed after passing a reporting point they may have assumed that the DC-9 was already behind them. Whatever, the circumstances, the two aircraft were closing with a speed through the air of around 920kt – at almost 1,100 mph faster than a rifle bullet—and there was almost no time to take avoiding action.
At JP550’s last call there were only 300 ft and a few seconds separating the flights. It was too late for Tasic to interrogate the computer for the DC-9s flight label as drastic action was required. BE476’s flight label incorrectly showed the Trident level at 335. Tasic surmised that if he could hold the DC-9 at its present level of 327 they would pass each other with 800 feet to spare. It would be close, but something had to be done. Turning the aircraft on to another heading would take more time and the action of holding JP550 level should have the desired effect. Tasic broke into Croatian and stammering, called back in a desperate attempt to avert a tragedy.

10.14:22 Tasic R/T: ‘. . . e . . . zadrzite se za sada na toj visini i javite prolazak Zagreba.’ (. . . e . . . hold yourself at that height and report passing Zagreb).

By now the DC-9 was climbing through 330.

10.14:27 JP 550 R/T: ‘Kojoj visini?’ (What height?)

Captain Krumpak immediately attempted to level the aircraft but for a few seconds its momentum carried it higher.

10.14:29 Tasic R/T: ‘No kojo ste sada u penjanju jer . . . e . . . imate avion pred vama na isn . . .? 335 sa leva na desno.’ (The height you are climbing through because . . . you have an aircraft in front of you at . . .? three three five from left to right).
10.14:38 JP550 R/T: ‘OK, ostajemo tocno 330.’(OK, we’ll remain precisely at three three zero).

Unfortunately, in spite of BE 476’s flight label on Tasic’s screen indicating the Trident cruising at 335, the aircraft was flying level at exactly 330. The DC-9 reached maximum altitude of 192ft above the Trident at 10.14:35 then drifted in a gentle descent towards the selected flight level of 330. For a few seconds, as the two aircraft sped headlong towards each other they were both precisely level. Tolerances in radio beacon calibration on altimeter mechanisms, slight variations in headings or wind fluctuations might just have combined to avoid a catastrophe, but it was not to be. At 10.14:41 the outer five metres of the DC-9’s left wing sliced through the Trident’s flight deck as the aircraft collided, killing the crew instantly. The stricken machines dropped out of control from the sky. The DC-9’s wing broke off and the left engine disintegrated with ingested debris. As the aircraft descended in flames the empennage detached. The Trident lost its rudder in the collision and dived steeply towards the ground. The cockpit voice recorder on the Yugoslav aircraft had not been working properly but the impact jolted it into action. As the DC-9 tumbled towards the ground it recorded Dusan Ivanus’ last words.

‘We are finished.’
‘Goodbye,’ he said, ‘goodbye.’

The two aircraft struck the ground about four miles apart near the town of Vrbovec lying about 16 miles northeast of Zagreb. All 54 passengers and nine crew aboard the Trident and 108 passengers and five crew aboard the DC-9 were killed. A total of 176 people lost their lives.
The first civilian airborne collision occurred on 7 April 1922 between a de Havilland 18 (G-EAWO) of Daimlar Airways and a Farman Goliath. The accident happened between Paris and Beauvais, the town associated with the R101, killing the two crew of the UK registered aircraft. History was also made over Zagreb that morning of 10 September 1976, for the mid-air collision between the Trident and the DC-9 proved to be the worst accident of that type on record. In the few seconds after the crash Tasic, in despair, called both aircraft in turn several times but to no avail. A Lufthansa Boeing 737, travelling eastbound on UB5 at level 290 towards Zagreb, was positioned only 15 miles behind the Trident. The co-pilot ‘saw the collision as a flash of lightning and afterwards, out of a ball of smoke, two aircraft falling towards the ground.’ The Boeing 737 commander, Captain Joe Kroese, reported the sighting to Erjavec, the middle sector controller.

10.15:40 Capt. Kroese R/T: ‘. . . e Zagreb! It is possible we have a mid-air collision in sight—we have two aircraft going down, well, almost below our position now.’ Erjavec was unable to understand and Captain Kroese had to repeat his words several times.
10.18:12 Capt. Kroese R/T: ‘it’s possible that the other aircraft ahead of us had a mid-air collision . . . er . . . just overhead Zagreb. We had two aircraft going down with a rapid rate of descent . . . and there was also smoke coming out.’

When the implication of what was being said dawned on Erjavec he glanced across at the upper sector controller. At his station sat a stunned Tasic, white faced with shock. Slowly he lifted the head-set from his ears and placed it on the console in front of him.

As the alarm was raised and rescue services were launched, operations ceased in the Zagreb control centre. Incoming aircraft to the airspace were diverted along other airways. The middle and upper sector controllers involved in the collision were suspended, and off-duty staff was summoned from their homes as replacements. Within an hour normal air traffic services were resumed. Soon police and officials arrived at the centre to interview those of the morning shift implicated in the catastrophe. Erjavec, the middle controller who had cleared JP 550 to climb; Pelin, his assistant, who had coordinated the flight with the upper sector; Hochberger, the upper controller who had vacated his station; Tepes, his replacement, who had been late for duty; Tasic, the upper sector assistant who had been left on his own for up to eight minutes at a critical phase; and Dajcic, the shift supervisor who had overall responsibility. Within hours they were placed under arrest. At Cologne Airport relatives and friends waited in vain for the arrival of the Index-Adria charter aircraft. Also kept waiting were the outbound group of tourists checked-in for the return flight to Split. All they were told was that the flight was delayed until 5 pm owing to a technical problem. Even 21/2 hours after the accident telephone callers were still being given the same facts. It wasn't until early evening that news of the crash was officially released to those waiting, but by then many had already heard of the tragedy from reporters assembling at the airport.

An initial accident inquiry established that the controllers had a case to answer, but the civil aviation authorities were released from the requirement to publish details. Instead, under Yugoslav law, the controllers also faced charges of criminal misconduct and arrangements were undertaken to place the accused on trial. Within two months all except Tasic were released. He, alone, remained in custody.
Seven months after the collision the hearing of the criminal case opened in the Zagreb District Court before Judge Branko Zmajevic at 8am on the morning of 11 April 1977. The court assembled in the same room where Tito had been tried as a communist in 1923. In the dock sat eight accused: the five controllers and the shift supervisor, plus two senior officials, the head of the Zagreb flight control service, Antere Delic, and the head of Zagreb district flight control, Milan Munjas. Deputy Public Prosecutor Slabodan Tatarac read the long list of charges of criminal negligence.
The trial was well conducted and followed Yugoslav law to the letter. The evidence was carefully presented and formal courtesies observed. The court, however, was not concerned to analyse the cause of the accident, which had been the purpose of the inquiry. It had assembled to apportion blame and to administer punishment. All defendants faced long prison sentences of up to 20 years if found guilty. As the trial progressed the atmosphere at times became heated and tempers flared. Claim and counter-claim followed accusation and counter-accusation. Statements by some of the accused contradicted or embellished evidence presented at the earlier accident inquiry and it became increasingly difficult to establish the facts. Pelin, the middle sector assistant, now claimed that Tasic had been shown JP550’s flight progress strip as early as 10.07 hrs when co-ordination of that aircraft’s climb first took place between the two. Pelin’s evidence was supported by Erjavec, the middle controller. None of this had been mentioned in the earlier court of inquiry. In turn Tasic presented fresh revelations. He now recalled that he had told Pelin JP550 could climb to 350 only if it could make level 310 by Kostajnica. If not, it would have to wait for climb until after Zagreb. Pelin denied this exchange. He countered by saying that during co-ordination of JP550’s climb he had also pointed out to Tasic Olympic 182 at Kostajnica level 330, which might have caused conflict with a clearance to climb. According to Pelin, Tasic replied that after they crossed, JP550 was cleared to climb.
The web of information became so tangled that any attempt at establishing the true facts seemed remote. The judge encouraged the accused to challenge each other with the facts but Tasic seldom appeared unscathed from these statements. He became increasingly isolated and fell ill under the strain. Proceedings had to be adjourned for a week.

In court, representing the prosecution was a British lawyer, Richard Weston. He attended in the capacity of attorney for the parents of one of the stewardesses killed aboard the Trident. His interest, of course, was that justice should be done, but as the proceedings continued he became genuinely concerned with the course of the trial. The evidence clearly demonstrated that the air traffic system was as much under scrutiny as the behaviour of the defendants and that changing the system would be far more beneficial than imprisoning the victims of the system. The accused, if found guilty, would be suitably disciplined by the civil aviation law. Weston, as an accredited representative, was entitled to be heard by the court. Before the judge’s summing-up he decided to exercise that right and, through a Yugoslav colleague reading a translation of his statement, he made an impassioned plea to the court for leniency. Although the controllers may be held responsible, they should not be sentenced to prison like common criminals. It was a most unusual comment from the prosecution benches, but one which was received with applause both within the court room and the world beyond.

On Monday 16 May 1977, the court assembled for the last time to hear the findings read by Judge Zmajevic. In short, of the eight accused, seven were found not guilty. Only Tasic was left to shoulder the blame. The evidence presented by Pelin and Erjavec regarding co-ordination of JP550’s climb was preferred to Tasic’s. He had made a number of contradictory statements. Tasic had not expressed surprise when given JP550’s flight progress strip 11/2 minutes before the collision. If he had not given permission to climb why did he not question the hand-over of the strip? On ascertaining the danger of the situation Tasic had neither issued to Captain Krumpak an order with authority nor stipulated an exact height. Had he done so expert witnesses had testified that there was sufficient time for the DC-9 to take avoiding action. The court concluded that Tasic was not overloaded and as such was in a position properly to assess the situation. He was found guilty of criminal negligence and was sentenced to seven years imprisonment.

An appeal was launched immediately and was supported by the sympathetic services of the International Federation of Air Traffic Controller’s Associations (IFATCA) who retained Weston on Tasic’s behalf. The Supreme Court reconsidered the case on 29 April 1978, and halved the jail sentence. IFATCA were determined to stop at nothing short of freedom for Tasic and organized a petition which was presented to Marshall Tito. The petition achieved the desired effect and on 29 November 1978, just over two years after the collision, Tasic was eventually released from detention.
Over the years pressure grew for re-examination of the facts which were first released after the original accident inquiry but were never made available for public scrutiny. In February 1982 the investigation into the collision was re-opened and in September of that year the results were finally published after a painstaking inquiry. Owing to conflicting statements, the report concluded, it was impossible to establish whether co-ordination of JP550’s climb had been effected. If co-ordination had been effected it had been conducted improperly by the middle controllers. It was impossible to establish whether Tasic approved JP550’s climb. The absence of Tasic’s partner at the upper station was against regulations and his departure remained unnoticed by the shift supervisor in spite of him being responsible for all staff. As a result, by any standard, Tasic had become overloaded as he tried to cope alone.”

Air Disasters by Stanley Stewart, Published by Ian Allen in