The secretive investigation into the crash of a SilkAir 737-300 in Indonesia shows up major cultural differences in dealing with accidents
Andrzej Jeziorski/JAKARTA
The thing about black holes, in astronomical terms, is that nothing ever comes out of them. Some people liken the controversial investigation into the crash in Indonesia of SilkAir flight MI185 in December 1997 to a mysterious cosmic phenomenon. One of these people is the chief investigator, Professor Oetarjo Diran, who also chairs the Jakarta-based Indonesian Air Accident Investigation Commission (AAIC).
When Diran talks about "the black hole", it is a tongue-in-cheek reference to the darkened office next to his, with a "no entrance" sign on the door. This room contains all the material gathered so far by the team of psychologists, psychiatrists and surgeons known as the Human Performance Factors Group, which is focusing on the most widely supported theory about the crash's cause.
Most reports call this theory pilot suicide, although, if proven, it would surely also be murder. There were, after all, 103 other people on board the aircraft when it plunged into the Musi River, about 55km (30nm) from Palembang in southern Sumatra, on 19 December, 1997. The aircraft was buried deep in the muddy riverbed, leaving investigators and search and rescue workers with fragments of wreckage swept away by the current, and a lot of dredging to do.
Seventeen months after the crash, Diran says the report of the investigation is only about 30% finished. In March, after much prevarication, the AAIC issued an interim report (Flight International, 21-27 April), which Diran says was produced under pressure from Singaporean authorities and victims' next of kin, but was not intended for public circulation - certainly not in Indonesia. Diran is concerned that the report fell into the media's hands, even though it reveals little that was not known a year earlier.
The bereaved families are angry and frustrated. In Singapore, a group of them recently petitioned prime minister Goh Chok Tong to press Indonesia to release more of the findings, which they believe are being suppressed. A similar petition is being drafted for presentation to the Indonesian Government once the post-election confusion settles. The group is also preparing to sue SilkAir, believing that this will allow them access to facts they believe are being withheld.
Now the AAIC has released a surprise statement saying it has obtained crucial financial records on the crashed aircraft's pilot which have been legally protected up to now. The revelation is all the more surprising because Diran showed little enthusiasm in pursuing these documents, which could provide overwhelming evidence to back the suicide theory. He says the pile of paperwork will take up to two months to sort - and even longer to analyse.
Diran, whose background is in engineering, claims not to know what human factors the investigators have unearthed in their interviews and investigations. "I don't want to know," he says, adding that they are the specialists in these matters. Diran believes that the group's findings - and all the investigation's findings to date - should remain under wraps until the conclusion.
But, given the state of the evidence to hand, Diran's secretiveness and his apparent squeamishness over the human factors issue, some people fear that the investigation's final report will contain no meaningful conclusions.
The known facts of the flight are as follows. The operator, SilkAir, is a wholly owned subsidiary of Singapore Airlines (SIA),which serves Asian destinations with Boeing 737s, Airbus A320s and Fokker 70s. The aircraft was a 737-300, registered 9V-TRF, powered by two CFM International CFM56 turbofans. At the time of its last flight, it was only 10 months old and had completed 2,200h and 1,400 cycles.
The aircraft took off from Jakarta's Soekarno Hatta International Airport at 15.23 local time, en route for Singapore. At the controls were Singaporean Capt Tsu Way Ming and First Officer Duncan Ward, a New Zealander. There were five cabin crew and 97 passengers on board.
Shortly after 16.00, the aircraft was cruising at 35,000ft (10,600m) with Ward at the controls and Tsu handling radio communications. But the final radio transmission from the aircraft was from Ward - suggesting that Tsu had either left the cockpit or was otherwise occupied. This last radio call was a routine communication with Indonesian air traffic control, and revealed no undue stress in the first officer's tone of voice.
Apparently unbeknown to the co-pilot, however, the cockpit voice recorder (CVR) had stopped working about 1min earlier. About 40s after Ward's transmission, the flight data recorder (FDR) also stopped.
At 16.13, the aircraft disappeared from radar screens. There was no distress call. Eyewitnesses saw it crash nose-first into the Musi. The aircraft, in Diran's words, was "pulverised" by the impact, having hit the river at transonic speed. Any hope of finding survivors was quickly extinguished. According to the AAIC, about 73% of the aircraft by weight was recovered by divers and dredging equipment.
Within weeks of the accident, the investigators reported that damage to the engines suggested that both had been running at close to full power on impact. Up to the moment the FDR deactivated, the aircraft's autothrottle is believed to have been engaged, which should have caused the engines to throttle back in the dive in response to overspeed conditions.
If the interpretation of the state of the engines was correct, then the autothrottle had been disengaged and the crew had taken manual control - as they might in an emergency. But to maintain full power in a dive defies logic. Safety specialists say one of the first "absolutely automatic" responses of any pilot in an unintentional dive is to "shut the throttles".
Diran says only that "the damage [to the engines] was severe". He says it has not yet been shown conclusively to what extent the state of the engines was caused by their throttle setting, and what is due to impact damage.
Fragments of the rudder and tailplane structure were found about 4km east of the crash site, sparking an Airworthiness Directive from the US Federal Aviation Administration requiring all operators of 737s delivered since 20 September 1995 to inspect the aircraft's empennages. It was later concluded that the empennage fragments were torn off the SilkAir 737 by aerodynamic loads as it approached Mach 1. Specialists say it is usually this part of an aircraft structure that fails first in such circumstances.
Theory after theory has been put forward and dismissed since then. Several aircraft on the same route at the time had asked for diversions around bad weather. Yet the AAIC says that reviews of meteorological data, including satellite imagery, have led to the conclusion that there was no evidence of weather severe enough to cause such an accident. There was no cumulonimbus activity; visual meteorological conditions applied; there were high cumulus clouds and light winds below 15kt (28km/h) up to 19,000ft - nothing out of the ordinary.
In its March interim report, the AAIC says: "The investigation found no indications that air traffic control, weather, maintenance, terrorist acts or hazardous material were factors."
Efforts are focused on two aspects, says the report. The first is "potential aircraft or aircraft systems anomalies that could have precipitated a loss of control". Here the report says the AAIC is checking whether "possible causes of past and recent 737 incidents involving sudden loss of control could have contributed to the descent from cruise altitude". Diran is particularly interested in the crashes of US Airways Flight 427 - a 737-300 - at Pittsburgh in 1994 and United Airlines Flight 585 - an older 737-200 - at Colorado Springs in 1991, both linked with uncommanded rudder deflections.
This link is being pursued with vigour by US lawyers attempting to sue Boeing on behalf of some of the victims' families. But there are key differences between the Pittsburgh and Colorado Springs cases and the case of the SilkAir 737 - and the most obvious is altitude.
United 585 left controlled flight at 1,200ft, hitting the ground 10s later. The crew of USAir 427 lost control while trying to recover from hitting the wake of a Boeing 727 at 6,000ft. Its descent lasted 24s - and even in this brief period, the crew managed to send a distress call. Both aircraft were flying slowly on approach - at about 158kt and 190kt respectively - close to, or below, the crossover speed below which rudder-induced roll can no longer be countered by aileron and spoiler inputs.
But the SilkAir flight was cruising at 35,000ft, comfortably above the crossover speed. Its descent reportedly lasted about 120s. The pilot had plenty of time to establish what was happening and recover, and the visual weather conditions would have helped his orientation.
Diran argues that an aircraft experiencing a rudder hardover would exceed the limits of its flight envelope in "4s", but Boeing aircraft have a reputation for robustness and have been known to survive stresses for which they were never intended. Boeing simulations show that even the Pittsburgh crash was recoverable - albeit under controlled circumstances, with pilots who knew what to expect.
In its report on the USAir 427 crash, the US National Transportation Safety Board (NTSB) says the "most likely" cause of the rudder hardover was a jam of the main rudder power control unit servo valve secondary slide. But in the SilkAir aircraft's case, this would still not account for the failure of the aircraft's voice and data recorders.
Some advocates of the uncommanded rudder hardover theory suggest it could have been triggered by chemicals - "blue water" - leaking from a toilet into the electrical and electronics bay. This, they say, could have caused the malfunction of the CVR and FDR and then the rudder malfunction. The likelihood that a chemical leak could cause such malfunctions within minutes of each other remains unclear. If it did happen, it would be unprecedented.
Blue water leaks have previously been reported in Boeing 737-300s, but none has been associated with a major system failure. Inspections of the SilkAir 737-300 fleet since the crash show no evidence of such problems, say sources close to the airline.
That leaves the final popular theory: that a crew member put the aircraft into its fatal dive deliberately, after deactivating the CVR and FDR via circuit breakers in a panel behind the pilot's seat. The difficulty is in establishing proof.
Although none of the CVR material has been released to the public and none of the investigating bodies involved is prepared to comment on its content, it is believed that conversation between the captain and first officer was relaxed until the recorder stopped working.
Sources close to the investigation say that in the final moments of the existing CVR transcript, Tsu is heard to say he is leaving the flightdeck, then there is a sound which could be Tsu's seat sliding back. Diran refuses to confirm this, but it seems to fit the hypothesis: reaching back to the circuit breaker panel would be awkward while sitting at the 737's controls.
While leaving the cockpit, however, it would be possible to deactivate the CVR by pulling its circuit breaker without attracting undue attention. With Tsu no longer in the cockpit, it would also make sense that the final communication with the ground, 1min later, was made by Ward.
When re-entering the cockpit, it would have been possible to deactivate the FDR and either disable the co-pilot or wait for a chance - such as Ward leaving the cockpit - to put the aircraft into its final dive, with no distress call required.
This type of suicide is rare in commercial aviation, but not unheard of. In 1982, a Japan Airlines McDonnell Douglas DC-8 co-pilot wrestled for control with his mentally ill captain who was trying to push the aircraft's nose down on approach to Tokyo's Haneda Airport. The resulting crash killed 24 of the 166 passengers on board. In 1994, a Royal Air Maroc ATR 42 crash in Morocco which killed all 44 aboard was attributed to pilot suicide, despite fierce protests from the Moroccan pilots' union.
The descent of the SilkAir 737 was monitored by radars in Jakarta and Singapore - with data coming to Jakarta from remote secondary surveillance radar (SSR) sites at Jakarta Halim Perdana Kusuma Airport, Semarang, Palembang and Natuna Island. The radar plot was replayed at Jakarta and the data were examined by Hughes Raytheon.
The radar plot was later used in simulator re-enactments of the flightpath, carried out in Indonesia and Singapore, and by Boeing in the USA. Diran is furious that some journals have quoted unnamed sources as saying that these simulations showed that the steep descent profile could only have resulted from deliberate action by a crew member.
"We did make some simulations-and of course we have computer simulations as well, but there are quite a lot of data," says Diran, adding that analysis of the data is not complete, and no conclusions can yet be drawn.
Diran says the front end of the aircraft was so badly fragmented by the impact that "nothing forward of the wing is recognisable", making it impossible to establish who was on the flightdeck or at the controls during the descent. Still, here lies a plausible modus operandi for suicide. The next thing to establish is a motive.
In March last year, the Civil Aviation Authority of Singapore (CAAS) confirmed that Tsu had been demoted from his position as an instructor with SilkAir in the months leading up to the crash, following complaints from colleagues. At the time, senior CAAS manager Tan Wee Lee said Tsu was unlikely to have been "overjoyed", but declined to elaborate.
The Wall Street Journal then reported that the demotion resulted from an incident in which Tsu had either switched off or erased a CVR recording of a flight during which he had made a bad approach, necessitating a go-around. This is corroborated by senior industry sources. The newspaper added that Tsu's co-pilot at the time - a New Zealander - registered a complaint with the airline and this incident scuppered Tsu's chances of a transfer to SilkAir's parent SIA, with its more lucrative pay scales.
It was later discovered that Tsu, a former air force pilot and one-time member of the Black Knights aerobatic team, had once been forced by mechanical problems to turn back from a McDonnell Douglas A-4 Skyhawk training mission that he was leading in the Philippines. The other four aircraft involved later crashed into a mountainside, with no survivors. The incident occurred on 19 December 1979 - 18 years to the day before the SilkAir crash.
Early last year, persistent rumours also began to circulate, originating from unnamed sources close to the investigation, that Tsu was heavily in debt from gambling and had suffered further big losses - some sources quoted as much as $2 million - when the Asian securities market crumbled. Similar anonymous sources say that in the days before the crash, the pilot had taken out several life insurance policies with a combined value of millions of dollars. Tsu left behind a widow and three children, but the family has declined all public comment.
Only the release of Tsu's financial records could turn these alleged debts and life insurance policies into any more than hearsay. Up to now, Diran has said the records are protected under Singaporean banking laws, preventing the investigators from examining them without a court order.
To the astonishment of crash investigators and safety specialists from other countries, Diran refused to do this, saying it was "not the accident investigator's role to get a court order", and stressing that the case "is not a criminal investigation".
Diran said going to court would be "beginning to apportion blame", which he understood to be a violation of Annex 13 of the Convention on International Civil Aviation, which states: "The sole objective of the investigation of an accident or incident shall be the prevention of accidents and incidents. It is not the purpose of this activity to apportion blame or liability."
Senior crash investigators from other countries counter that "any investigation must pursue all avenues" to gather relevant information. Although apportioning blame is not the intention of an accident investigation, it is hard to imagine a conclusive investigation that does not do so to some extent. Diran's interpretation of this clause of Annex 13 seems unusually extreme.
His interpretation seems to arise from a fear of antagonising anybody while he struggles to establish an air safety culture in Indonesia. Diran is pushing for the formation of an NTSB-type safety body (Flight International, 17-23 March), and wants to remain on good terms with other aviation authorities, airlines and manufacturers while doing so.
"In Asia in general, in Indonesia in general, loss of face is important," he says. "I am not going to make enemies from the beginning." Diran denies that he is under any pressure, either from his own government or from Singapore, to keep a lid on the investigation's findings. "[I am] responsible only to my conscience," he says.
Suicide theory
Tsu's state of mind in the run-up to the crash has been a focus of the investigation since soon after the incident, and his financial records are critical evidence for the suicide theory.
Now, to the delight of the bereaved families, these records have been released without either a court order or any obvious effort from Diran.
"We got them by special permission from the [pilot's] estate," he says. The pilot's family had been approached by the Singaporean authorities after pressure was exerted on Singapore by governments of other countries - such as New Zealand - whose citizens had been killed in the crash, says Diran.
He insists there was "no pressure from the authorities" on the pilot's family to release the documents - despite the fact that evidence of suicide would automatically jeopardise any life insurance payouts.
Representatives of the Singaporean crash victims' families say they still fear the investigation may be deliberately dragged beyond the two-year legal deadline for compensation claims. They say suing SilkAir could give them independent legal access to the evidence needed to pursue those claims. But they must first reach an agreement with US lawyers pursuing Boeing, who fear that action against SilkAir could compromise their own litigation.
For its part, SilkAir has increased its compensation offer to the next-of-kin from the $75,000 specified by the Warsaw Convention to $135,000. On lawyers' advice, few of the families have accepted. Lawyers on the US cases also complain that the investigating bodies involved, including Boeing, the NTSB and the CAAS, know much more than they are revealing.
Under Annex 13 of the Convention on International Civil Aviation, participants in an investigation cannot reveal details "without the consent of the state conducting the investigation". But in a December 1998 letter to Diran, the Chicago-based Nolan Law Group wrote: "There is nothing in Annex 13 to prevent you from giving your consent. Your unwillingness to allow disclosure is now obstructing and infringing upon the rights of the next-of-kin to move forward with their claims for compensation."
This holds little sway with Diran, however. "My aim is to ease the burden [of the relatives' grief]," he says, "but I am not going to help them get compensation. I know there is a deadline, but that's not a criterion for pushing harder [with the investigation]. Annex 13 says that my investigation will not be used for litigation."
Diran says other countries cannot expect the same degree of openness from an Indonesian investigation as they can from, say, the NTSB. "We have our own laws for investigation, but people in the USA are demanding the same amount of transparency and public hearings [as in their country]. My country is not like that."
That drew an angry response from family representatives, who say: "The investigation is either incompetent, or they are dragging their feet for some reason." They hope their appeals to the Singaporean and Indonesian Governments will increase pressure on the AAIC to release the findings to date.
But the authorities in these countries are well known for their lack of enthusiasm for transparency and controversy. The families are doing their best to raise the profile of a case which they feel some would prefer buried.
Source: Flight International
