Opposites Attract: Seeking Closure Down In Brazil
As the investigation into the fatal midair collision of the GOL 737 and the Embraer Legacy over Brazil’s Mato Grosso jungle proceeds, definite facts are beginning to emerge.
Mixed in with these facts are more nebulous assertions. Because these assertions are being published by reputable Brazilian periodicals, it’s worth examining them as well, in an attempt to get the complete picture of what went so horribly wrong.
GOL airlines Flight 1907, a new Boeing 737-800, collided with a Brazilian-made Legacy jet over the Amazon jungle on Sept. 29. The damaged Legacy landed safely at a nearby air force base, but the larger airliner crashed in dense jungle, killing all 149 passengers and six crew members.
There are two schools of thought on whether the Legacy had the right to stay at Flight Level 370 after losing communications with its controlling Center. One school believes that, notwithstanding what’s written in the Filed Flight Plan (FPL), the FPL by itself is no authorization to change levels in controlled airspace — i.e., one should always maintain the last cleared level after a nocomm situation arises, regardless if it’s inappropriate for the new track. They assert you should let ATC interrogate your transponder, read your mode Charlie height response and act accordingly to ensure separation
The other school claims that, as Brazil is an ICAO signatory, pilots should always comply with the ICAO rules (and not the FAA or Hong Kong rules for nocomm). And those rules say:
ICAO Annex 2: 3.6.5.2 Communication Failure
If a communication failure precludes compliance with 3.6.5.1, the aircraft shall comply with the voice communication failure procedures of ICAO Annex 10, Volume II, and with such of the following procedures as are appropriate. (Italics are inserted by ASW, for emphasis.)
The aircraft shall attempt to establish communications with the appropriate air traffic control unit using all other available means.
3.6.5.2.1 If in visual meteorological conditions (VMC), the aircraft shall:
a) continue to fly in visual meteorological conditions and at the nearest suitable aerodrome; and report its arrival by the most expeditious means to the appropriate air traffic control unit;
b) if considered advisable, complete an IFR flight in accordance with 3.6.5.2.2.
3.6.5.2.2 If in instrument meteorological conditions (IMC) or when the pilot of an IFR flight considers it inadvisable to complete the flight in accordance with 3.6.5.2.1 a), the aircraft shall:
a) unless otherwise prescribed on the basis of regional air navigation agreement, in airspace where radar is not used in the provision of air traffic control, maintain the last assigned speed and level, or minimum flight altitude if higher, for a period of 20 minutes following the aircraft’s failure to report its position over a compulsory reporting point and thereafter adjust level and speed in accordance with the filed flight plan;
b) in airspace where radar is used in the provision of air traffic control, maintain the last assigned speed and level, or minimum flight altitude if higher, for a period of 7 minutes following:
1) the time the last assigned level or minimum flight altitude is reached; or
2) the time the transponder is set to Code 7600; or
3) the aircraft’s failure to report its position over a compulsory reporting point; whichever is later, and thereafter adjust level and speed in accordance with the filed flight plan.
Applicability Of The Rules
Ostensibly it’s either para 3.6.5.2.1(a) because they were in clear VMC cloudless conditions or conceivably 3.6.5.2.2(b) 3 for an IFR flight. However, under the latter clause, the crew’s culpability for not waiting 7 minutes and then dropping to the correct FL360 for NW bound traffic on airway UZ6 (i.e., per flight-plan) is unclear. They could claim that they were in compliance with the VMC provisions. They could also say that, at the time of the collision, they were still trying to achieve comms and hadn’t concluded that “they” were NOCOMM (and so should be squawking 7600). That decision to squawk the transponder code for NOCOMM relates to “whichever is later” (of the three provisos).
If the Legacy crew had tried to relay through another aircraft, that would be a good defense for them, demonstrating that they were still trying to establish comms. It would prove something that all local pilots were aware of anyway, which is that between Manaus and Brasilia, there are a number of communications blind-spots and that it was the two Control Centers (Cindactas 1 & 4) themselves that were incommunicado.
A lawyer could have a field day with the intricacies of the situation in which they’d suddenly found themselves. Not the least of these local peccadilloes is the fact that the Brazilian ATC centers tend to use Portuguese when speaking to local aircraft and English to others. That is a classic situation that worldwide tends to rob all (not just “other”) pilots on frequency of situational awareness. It also makes it harder to achieve a good comms relay. Consequently, could the ATC system’s hardware or personnel been more at fault?
Brazil’s ATC
Brazil upgraded its airways system as part of a $1.4 billion array of ground and air sensors designed to allow radar surveillance of airways over all of Brazil, which features a vast trackless expanse of forests, mountains and rivers. It was a Raytheon project completed in July 2005. Many local pilots have now said that radio black-spots still exist, some of which last for minutes at a time.
“At normal flight levels for commercial aircraft we have full radar and radio communications everywhere,” objected Brig. Gen. Alvaro Pinheiro Costa. He oversaw the creation of the Amazon’s radar coverage. He claims that calibration by an aircraft that flew the fatal route shortly after the crash confirmed that radio and navigational aids along the way were operating normally.
However, a number of commercial airline pilots, both American and Brazilian, say the air traffic control system over the remote Amazon River basin remains riddled with communication gaps.
Elnio Borges, 53, a pilot who has flown jets for Brazil’s Varig Airlines since 1980, says: “The guys who fly those routes expect to lose contact between Brasilia and northwest Brazil. It’s amazing that they are now claiming that these spots are not there.” The collision happened above a remote rainforest near the Chachimbo Mountains, an area that according to Flight International is near the limit of radar coverage between Manaus and Brasilia. The area is one in which radio communication can be especially unreliable, according to pilots who regularly fly the region.
The American pilots, Joseph Lepore and Jan Paladino, told Brazilian investigators in their depositions that they had tried but failed to make radio contact with Manaus in the 10 minutes before the two aircraft met head on at around 900 knots. Whether they went back to their last frequency is unknown. Whether they might have had (or been given) the wrong VHF frequency to “switch to” is not known. Usually that glitch can be sorted out by asking “in the blind” for any other traffic on that frequency to respond.
Where individual states decide to vary from ICAO procedures, these differences are laid down in a standard Annex to their own Aeronautical Information Publications (AIPs). Long-haul pilots wouldn’t be familiar with all the nuances of each and every country that they fly through or over — and that in itself is a trap. DAC, on their Brazilian Civil Aeronautics website, doesn’t help by only publishing theirs in Portuguese.
Here’s a good translation of the applicable section (and vive le difference). We would hope that messrs LePore and Paladino are aware of the important distinctions therein, in allotting responsibility for establishing positive separation for nocomm traffic.
ICA 100-12
RULES OF THE AIR AND SERVICES OF AIR TRAFFIC [2006]
(Brazilian Rules)
7.14 GROUND AND AIR COMMUNICATIONS FAILURES
7.14.1 when ATC units cannot maintain two-way communications with an aircraft in flight, they should take the following measures:
a) verify the aircraft can receive the unit’s transmissions by requesting in the blind that it execute specific maneuvers that can be observed on radar or that it should set, if possible, a discrete transponder code with the purpose of confirming the reception of ATC’s messages; and
b) if the aircraft fails to respond, the controller should maintain positive separation between the comms failure aircraft and all others, whilst assuming that the aircraft will adopt the established procedures for communications failure.
Source: tinyurl.com/yd6kl3 (in Portuguese). Theoretically, if a radar controller is holding an aircraft’s primary “paint”, he should also have, adjacent to that blip, its data block moving along with its secondary radar return. However, to some extent that can depend upon aspect (plus transponder antenna positioning) and range. If the line-of-sight between the Legacy and Manaus (Cindacta 4) was blocked, conceivably the signal could have been intermittent at longer ranges, but the nearby Cindacta 1 (Brasilia) should still have been recording their squawk. Honeywell is now claiming that the Legacy’s transponder (by its serial number) should not have been prone to going to STBY (see Oct. 9 issue of ASW).
Assuming that the Legacy pilots did not switch off their transponder, that now supposedly inapplicable standby flaw is still not the only plausible answer to the question of why TCAS was not working for them. It is possible, but unlikely, that the pilots were checking out some of the electrical bus change-over features of the Legacy and thus were unintentionally disabling their squawk for a few minutes at a time, as they switched busses in and out.
When VMC Prevails..
VMC prevailed at the time. Precisely, yet unsurprisingly, the aircraft hit each other anyway. We all know that the first duty in any lost comm procedure is to maintain VMC if practical and land.
Unfortunately, the key statement here is “if practical”. Consider that in these flight levels the closure rates are about 900 knots, not to mention the blind spot that covers approximately 90 percent of the surrounding environment. So, is using a “see and avoid” solution ever really “practical”? A timely sighting of a looming “other party” coming head- on is also highly unlikely. Furthermore, consider this: virtually every Mid Air Collision on record has occurred in VMC, with few exceptions.
In fact, the event that caused the creation of the U.S. domestic air traffic control program was a midair collision in visual conditions between two transcontinental airliners over the Grand Canyon. “See and avoid” really depends upon “see and be seen”. It takes two to tango, and there’s no need to seek closure. Modern avionics accuracy is perversely a highly noxious form of mutual magnetism.
Viewing all participants objectively, it was as if they were all “seeking closure” through inaction. That final fatal attraction over the Mato Grosso stemmed directly from an over-reliance upon TCAS. To quote Strother Martin’s famous movie line: “What we have here is a failure to communicate.”
Indeed, the failure was at all levels. However, the real legacy of this accident might be a fresh look at mandated offset tracking on airways (see http://www.iasa.com.au/offset.htm). It’s now accurate to say that accuracy can be the real killer. The “Big Sky” theory of putatively random misses is a thing of the past.