IATA Group has tracked data since 2006

[Don Hudson]

While this is good news (about which I'm not unhappy to have been wrong in believing the trend would increase, not decrease), the most interesting comment comes from Perry Flint, "IATA's Washington-based spokesman" who says, "The 2011 rate of 0.34 accidents per million flights is the lowest since World War II, when the modern airline industry began, . . .This is a long term trend. 2011 is not some kind of anomaly." From the article, "The number of deaths in airline accidents globally is at the lowest level since 2006, the first year for which IATA reported data in that category on its website. "

Mr. Perry's optimism is understandable given 2011's statistics and in this business, unlike the stock market, sometimes history and even those "forward-looking statements" do mean something but I would like to know from where does his and IATA's certainty come? IATA has been tracking accidents "since 2006" and claims a trend which is not an anomaly and which will not alter?

A mild case of spin perhaps? No one in this business has the gift of prophecy, especially after only five years so it logically can't be otherwise. So we don't know what the statement really means although we understand that it is pointing to a very good performance by the industry for 2011 and that is a reasonable thing to do.

The Statistical Summary of Commercial Jet Airplane Accidents Worldwide Operations, 1959 - 2010 is from 1959 to the present and may provide a more detailed, meaningful picture.

IATA is the international lobby group for air carriers so theirs is a vested interest just like other organizations who may portray and prod in different directions. I've always preferred the accident investigation organizations like the NTSB and the TSB because their vested interests are about "finding things out" and such activities aren't fundamentally economically or politically-driven. While aspects and causal factors can be emphasized or de-emphasized, (the BEA's Final Report on the AF447 accident is going to be very interesting), you can't put 'spin' on an accident investigation in the same way that the lobby process works in the back-rooms towards political or economic ends. Not that "back-room work" is always bad...often its a good way to accomplish something but such work intends to keep the process of sausage-making from delicate eyes and minds while the investigation process must go where the evidence leads regardless of who it may touch.

So Mr. Perry's remarks must necessarily be taken with the usual grain of salt, which is not how one usually takes the results from organizations like the NTSB and the TSB here in Canada. One hopes he is right though. There are good reasons cited in the article, not unknown to most here, for optimism. But I'm suspending judgement in favour of curiosity if only because the worst thing one can do in this business is pronounce on how safe one's airplane, one's airline, one's industry is. In this business, genuine comfort with the way things are is inappropriate.

In the meantime the actual numbers are certainly gratifying. Though unrecognized by those doing the pronouncing in Washington, the engineers working with such systems mentioned in the article should take satisfaction in knowing that they had a hand in creating such positive outcomes for our industry. When I think of this aspect of aviation I always return to one man, Don Bateman. Mr. Bateman, a Canadian and a graduate of the University of Saskatchewan (BE, 1956) was an engineer at Honeywell when he invented the ground proximity warning system and a host of other safety systems. Some of the early tests Bateman and his team did in proving the system were against Mt. Slesse, just east of Vancouver, where a CFIT occurred to a TCA Northstar 56 years ago this month.

Don

Group that has tracked data since 2006 sees 'long-term trend'

The global airline accident rate is on pace this year to reach the lowest recorded, according to a trade group.

The rate of accidents serious enough to destroy an airplane was, through November, 52 percent lower than the average for the previous five full years, according to statistics compiled by the International Air Transport Association.

The 2011 rate of 0.34 accidents per million flights is the lowest since World War II, when the modern airline industry began, said Perry Flint, IATA's Washington-based spokesman.

"This is a long-term trend," Flint said. "2011 is not some kind of anomaly."

The number of deaths in airline accidents globally is at the lowest level since 2006, the first year for which IATA reported data in that category on its website. Through November, 486 people had died in air crashes, compared with the previous low of 502 in 2008.

IATA represents airlines, including Aer Lingus in Ireland, Air France and US Airways Inc.

One year of data should be approached with caution because trends in aviation safety take many years to develop, said Kevin Darcy, a director of RTI Forensics in San Francisco and former chief accident investigator for Boeing Co. RTI performs engineering consulting on aviation and marine accidents.

Still, technological improvements in the manufacture of aircraft and safety devices introduced the past two decades have almost eliminated some types of accidents, Darcy said.

Cockpit databases that track a plane's location and warn pilots when they get too close to mountaintops or other obstructions have made rare what were once causes of many accidents, Darcy said.

Other safety systems have helped reduce in-air collisions and wind-shear crashes, he said.

Africa remains the most dangerous region in which to fly, according to the IATA data. There were 3.93 serious accidents per million flights through November. That level was 29 percent less than the average for the previous five years.

Europe and North Asia had no serious accidents this year, according to IATA.

There has been one serious accident in about 10 million flights in North America, according to the group's data. That Aug. 20 crash of a First Air chartered Boeing 737-200 in Resolute Bay, Canada, killed 12 people, according to Aviation Safety Network, a website that reports accidents. First Air is owned by 9,000 Inuit of northern Quebec.

IATA's data track crashes involving scheduled and nonscheduled carriers around the world. It doesn't count terrorist attacks or accidents involving private aircraft.

The IATA definition of a serious accident also doesn't include some occurrences aviation investigators consider dangerous. For example, a Southwest Airlines Boeing 737 that lost a 5-foot section of fuselage skin April 1 over Arizona isn't included because the damage wasn't sufficient to destroy the jet.

The National Transportation Safety Board considered the episode serious enough to conduct a public hearing. The tear was traced to a manufacturing error, the NTSB has said.

The most common type of serious accident occurs when an airplane goes off a runway during takeoff or landing, according to a Dec. 7 presentation by Gunther Matschnigg, IATA's senior vice president for safety, operations and infrastructure.

Those accidents accounted for 23 percent of serious crashes, Matschnigg said.

"IATA understands that there is still a lot of work that needs to be done in this area," he said.

http://articles.chicagotribune.com/2011-12-29/business/ct-biz-1229-airline-accident-rate-20111229_1_accidents-iata-aviation-safety-network

[Specs]

Tacoma Narrows Bridge

Fukishima

Titanic

Challenger

Hindenberg

I bet all of those represented the leading edge of technology in their time and had trend monitoring been conducted, the future would have looked pretty rosy at the time.

Trend monitoring is a great analytical tool until it isn't. It says what was - how we've managed to date - but means little in terms of the future in an evolving environment.

With an almost exponential increase in the # of operators worldwide and the downward pressure on salaries, training and qualifications, etc, - ask yourself just how comfortabe you are relying on past trends in today's world of SMS, MCPLs, Biofuels, Automation, new materials, new environments (GNSS) etc.

IMO, we're always pushing the limits of the envelope somewhere and we're never far enough away from a major catastrophe to stop being vigilant.

[Don Hudson]

The Challenger/Columbia accidents and Fukushima share much in common beginning with the unwillingness to listen to experts whose thinking went counter to the popular view (spin?). The exposure of the Fukushima Nuclear Plant to a single-point failure due to the unprotected primary back-up power plant which failed completely an hour after the earthquake after the installation was flooded, was discussed by scientists and engineers and brought to the attention of TEPCO which dismissed the concerns citing historical records that an earthquake the size of the one that destroyed the plant was not only improbable but not possible given seismic history - so sayeth TEPCO. Now, the Japanese government is forced into bailing TEPCO out by taking it over "temporarily" because the privatized company cannot pay the billions in compensation without going out of business. Once again, so much for private enterprise doing the right thing when the 's' hits the fan.

The issues are much larger however, which take this far beyond individual accidents such as Challenger and Columbia. Private enterprise by its very nature must put a blush on everything it does and must deny the seamier sides of enterprise with happy-chat, otherwise shareholder value and profit suffer. There is something contradictory, amoral and sinister about such a conflation of goals and responsibilities, isn't there?

It's not "private enterprise" I argue against, (I'm certainly not in favour of more unelected bureacracy), but the fact that when it comes time to accept responsibility for the activities and outcomes they are engaged in, they run behind the skirts of the public purse via government, the courts, even public opinion and social media to plead their special version of the socialization of risk and failure.

Business would enjoy a far better reputation if it actually behaved maturely as we expect our children to behave. But planning and acting responsibly so that "the commons"...the public itself is not placed at undue risk in the name of insitutional denial, shareholder value and profit is a minimum qualifier in the invitation to a friendlier public reception. At the moment, who really trusts private enterprise to do the right thing? Government at least must feign a response, but its record of sitting on the side of the people it governs is not good either, not, at least, in North America.

All we need to do is take a look at how two companies behaved during the massive environmental disaster that resulted from the Deepwater Horizon fatal explosion and fire, Transocean, the owner of the Deepwater Horizon, and BP which continued stance of denial and irresponsibility was only exceeded by its CEO's stupidity and inability to comprehend what had occurred and why. These are "normal" accidents in the sense that their origins were known and understood beforehand but for reasons only acknowledged after each accident, were not addressed.

Charles Perrow's book, "The Next Catastrophe" discusses these topics, and the list is much, much longer than these three present-day accidents and environmental disasters.

Don

[J.O.]

The IATA article carries a hint of hubris that smart safety managers would never entertain. It makes me think of a comment I heard from a pilot at a party last night. We were talking about his perceptions of SMS and the "dumbing down" of oversight by the regulator. When I suggested that operators who do a good job of managing risk deserved to be less under the regulatory microscope, he responded in effect that a record of operating safely is much like a financial prospectus - past results are not a guarantee of future performance. Tough point to argue with.

[Don Hudson]

Hi J.O.;

Re >> . . . he responded in effect that a record of operating safely is much like a financial prospectus - past results are not a guarantee of future performance. Tough point to argue with. <<

(Not saying anything here that you don't know and can recite in your sleep, just discussing the point made by your friend).

It's a fascinating comparison!

For me, it only appears similar but " financial prospecti " are rosy-glasses documents not written by SEC bureaucrats who we would expect to be dispassionate and even skeptical but, to be blunt but not unkind, by company pitch-men for the stock even though their legally-required caveat-emptor cautions may be present. The institution itself is very much casino-like in origin and subsequent historical performance, (and even that statement is worth examining more!)

Aviation is fundamentally less speculative and I think it is a bit more reasonable to make the statement that past history can be an indicator of a carrier's potential for an incident/accident. As an example of what is meant by this, we could compare Jetsgo's history, even before their YYC runway excursion, with any other carrier in Canada at the time and even today and decide "who's next?" (To complicate matters, most of us make a 'gut-feel' assessment on such questions which might not be unreasonable, but to keep the question open for just a moment, we could argue statistically that those with more flights may have a greater probability of an accident, or we could argue statistically about incidents/events per million hours, per thousand landings etc and come up with the opposite view - our 'gut-feel'. For a really interesting take on this kind of question (gut-feel or logical reasoning), see Jeremy Clifton's profile of Daniel Kahneman on the IMF website or take Kahneman's "test" on thinking (which he wants everyone to fail !)

The term, "stochastic" (which see) is helpful to describe such a phenomenon because an accident can occur anywhere, even in the very best, risk-averse well-run organizations. I've used the example of a person crossing a freeway: we all know that it is possible and if careful, probable that the person will make it across 'n' times but we also know that at some point the person is not going to make it across but we can never say on which crossing the person is going to be hit even if we had all the necessary information such as traffic flow, which lanes, visibility, the capacities of the person and the driver who eventually...well, you get the (morbid) idea. And the person may not be hit at all. The permutations of all possible factors which wind their way in time and over unpredictable terrain like trickles forming brooks forming tributaries forming rivers forming oceans, bury the sources of causal pathways and the slightest variations in "terrain" alter such courses indeterminately and we are left with hindsight which explains things in an "aha..!" fashion only afterwards.

We cannot say anything for sure, which makes me believe that IATA doesn't really understand flight safety.

A diversion...

Lee Clarke has written a really neat book about "possibilistic" (worst case) thinking, (as compared to "probabilistic" thinking). He says, in part:

Worst case thinking is very different from the modern approach to risk. The modern approach is based on probabilistic thinking - what's the likelikhood that the nuclear plant will melt down? - and is often used to justify dangerous systems. Worst case thinking is possibilistic thinking - what happens if the nuclear plant has a really bad day? - and can be more progressive.

The question for aviation is one of frequency and cost because the question of "worst case" can be asked of every flight. In examining Fukushima, the cost of ensuring a secure, reliable first-response power supply for the plant to continue its shut-down processes with is essentially one-time, as was a proper installation of a working blow-out preventer for the Deepwater Horizon. The notion behind "possibilism" is the same kind of thinking we do when writing out wills...What does the worst case look like, who is hurt and how can that best be prevented? In aviation, SMS and the various programs which work under SMS's umbrella are largely built upon probabilistic thinking. In aviation then, "worst case thinking" belongs upstream of SMS programs. It should already have been done and this is in fact generally the case.

If I may offer a thought: - If there were SMS Programs for the speculation economy and the stock market we might not have as many crashes.

Don

[J.O.]

David Learmount takes a different look at the issue and what the future may bring.

http://www.flightglobal.com/news/articles/in-focus-airlines-run-up-a-safety-debt-365509/

[Specs]

David Learmount takes a different look at the issue and what the future may bring.

http://www.flightglo...ty-debt-365509/

A different look? Really? Those all seem fairly familair issues to most on here I would think. Makes me even suspect David Learmont is a lurker on AEF.

(hmm? members map shows no pins on England)

[J.O.]

You're right, those of us inside the industry have known those truths for some time now. I guess what I meant was that his take was different from the one taken by IATA, which was all sunshine and roses.

[Don Hudson]

David Learmount takes a different look at the issue and what the future may bring.

http://www.flightglo...ty-debt-365509/

The last 3 paragraphs bring the point home.

For those reading this who may not be aware, airline pilots, particularly those doing long-haul on large transports, typically actually fly the airplane about 3 to 5 hours per year.

Here's why: We fly perhaps five minutes at and after takeoff and less if the SID (Standard Instrument Departure) is compicated. The autoflight is engaged for the climb, cruise, descent and approach until about 400ft above ground at destination. That's about another 30 seconds of manual flight or, if one disengages at the Outer Marker/Final Approach Fix one may hand-fly for two minutes.

Long haul pilots will typically do perhaps three, at the most four takeoffs and landings per month. Assuming one flies "the maximum" outlined above of seven minutes that's roughly 28 minutes or at the very most, roughly a half-hour every month of actual hands-on piloting of the machine. Given equipment changes, reserve flying, vacation, displacements for training and occasional illness the hypothetical six hours of hand-flying reduces to three to five hours of actual time.

This is no longer just an individual airline deciding that hand-flying will not be encouraged any more. Squeezing more airplanes into the same airspace requires navigation and altitude accuracies that are available only through automation. The aviation transportation system itself has already implemented RNP 0.1 standards where I imagine that hand-flying will be against the law. From the RNP Wikipedia article:

"An RNP system utilises its navigation sensors, system architecture and modes of operation to satisfy the RNP navigation specification requirements. It must perform the integrity and reasonableness checks of the sensors and data, and may provide a means to deselect specific types of navigation aids to prevent reversion to an inadequate sensor. RNP requirements may limit the modes of operation of the aircraft, e.g. for low RNP, where flight technical error (FTE) is a significant factor, manual flight by the crew may not be allowed. Dual system/sensor installations may also be required depending on the intended operation or need.

"An RNAV system capable of achieving the performance requirements of an RNP specification is referred to as an RNP system. Because specific performance requirements are defined for each navigation specification, an aircraft approved for a RNP specification is not automatically approved for all RNAV specifications. Similarly, an aircraft approved for an RNP or RNAV specification having stringent accuracy requirements is not automatically approved for a navigation specification having a less stringent accuracy requirement."

ATC and those designing SIDs, STARs and Approaches plan routings, altitudes and speeds on the accuracy that automation provides and it is a practical requirement and no longer a choice to engage the autoflight system. Hand-flying a SID out of London or Frankfurt just isn't done. One can more easily fly the descent and approach manually but in heavy traffic areas the workload is said to increase. I say "said to increase" because quite frankly if one knows the airplane well, the workload does not increase. It only increases if one has become so reliant on the autoflight that assisting the pilot flying actually "out of role" and is therefore an increase in workload. The workload for the pilot-not-flying is not technically demanding, it is psychologically demanding and the industry is not dealing with this essentially-human factors issue. The workload for the pilot flying may or may not be higher depending upon skill level.

I think this is where the reduced skill levels, in terms of performing as an air crew, are coming from.

It boils down to money and the impatient demands for profit and shareholder value but primarily what I see is a significant decrease in the understanding of "aviation" and what it still takes to fly complex airliners in a complex transportation system with sustained safety.

Many of us here who are commenting on this trend and who have grown up in a system that pre-dates the advent of what I call "deep automation" are able to place automation appropriately...waaayy down below, in steerage where it belongs because, like computers, automation is a high-speed idiot without perception, values or goals. But seen, understood and used appropriately is of enormous benefit. But it takes effort and careful training to so place such an enticing bit of technology where it truly belongs. I think the infatuation with automation borders on dependency in a number of ways which are slowly edging out awareness of the shortcomings of the absence of "hands-on", and not only in aviation.

From the Flight Global article quoted by J.O.:

. . . .

"What has most affected the nature of pilots' work is the influence of low-cost carriers, which has brought radical change in many airlines' relationships with flightcrew. But what has most changed an airline's crew recruiting and management is the decline of the military as a provider of pilot skills.

Meanwhile, there has been a loss of pilot exposure to ­anything other than pre-packaged flight planning, followed by automated flight on the line. When ­circumstances are unusual, non-standard, or not automated, a resulting lack of pilot resilience has been leading to fatal loss of control (LOC) accidents, making LOC the biggest killer accident category this century - taking over from controlled flight into terrain in the last.

"This fact is acknowledged by industry bodies like the International Air Transport Association (IATA) and the International Civil Aviation Organisations (ICAO), respectively IATA's training and qualification initiative (ITQI) and ICAO's next generation aviation professionals (NGAP). So the carriers cannot say they have not been warned, but these efforts have not been translating into action at airline level.

"Just as a reminder, the number of fatalities caused by airline accidents in the 1980s was about 1,100 annually, whereas the numbers now are less than 800 a year despite the fact that the revenue passenger kilometres flown now are three times what they were then. The industry could revert to the bad old days, but for a different reason: now the aircraft are better, but the skills to operate them are degrading."