Plenty of media coverage was devoted to the Southwest Airlines meltdown that occurred in the second half of this past December. The beleaguered air carrier canceled 16,700 flights over the holiday season at an estimated cost of between $725 and $825 million (around half to cover refunds, the rest for reimbursements). This total equals the amount that Southwest earned in the first nine months of the year. The cancelations equaled the total for the preceding 10 months (which had already drawn White House ire due to poor, post-COVID performance). The opaque cost of bad press will certainly push the burden much higher. The exact number of affected passengers has not been disclosed, but a fair estimate has it at seven figures. As much as a
half-a-percent of the total population of the United States was impacted.
The triggering incident was related to a “bomb cyclone” of freezing weather and wind. The severe conditions saturated large swaths of the United States, creating chaos for national air carriers. Yet for Southwest, the disruption extended well past the worst of the weather. There are a few reasons for this. First, Southwest utilizes a point-to-point network (as opposed to the more traditional hub-and-spoke of competitors). This model can be cumbersome when confronted with scheduling discontinuities. Air crews are spread randomly around the country, making recovery more difficult when cancellations occur en masse. Secondly – and more relevant to the ensuing meltdown – Southwest utilizes an antiquated network to reassign displaced flight crews. In a world increasingly reliant on automated solutions, aging technology has proven to be an Achilles heel.
The winter storm began on December 21. But the worst of the cancellations did not occur until the 26 when over 50 percent of Southwest flights were canceled. The technological meltdown drew a harsh response from the Department of Transportation. This rapidly evolved into an egg-on-face moment for the FAA. A week and a half after Southwest had unraveled the glitch, a nationwide ground stop was issued due to a failure in the government’s Notice to Air Missions (NOTAM) network. The ground stop wreaked havoc on early morning departures. It resulted in just shy of 7,000 delays and 1,100 cancelations. A nationwide ground stop is a rare event reserved for extraordinary circumstances. The last time it happened was in 2001.
Initially, the idea circulated that the NOTAM failure was the result of a cyberattack, but the final cause was much more benign. An IT contractor had unintentionally deleted files while “working to correct synchronization between [primary] and backup databases.” Media experts soon postulated that antiquated IT systems had been the culprit. Some questioned the robustness of the NOTAM delivery mechanism. To be fair, the current system has decades of reliable operation under its belt. Still, the fiasco points to the obvious fact that a sole source failure was enough to paralyze domestic air travel – even if only for a handful of hours (most of which occurred in the middle of the night).
If you are into conspiracy theories, you might be inclined to think that the FAA did it on purpose. The 5-year budgetary process for the agency is up for reauthorization this year. Immediate talk by DOT representatives focused on the budget as a solution to the dilemma. The current NOTAM system is three decades old. Renewal is not slated to occur for another several years. On top of this, the current NOTAM system already has its detractors. Designed primarily to inform pilots of non-published restrictions in the national aerospace system, it often serves as a dumping house of operationally marginal data. Three full pages of obstacle locations (listed via long-form coordinates) is the sort of information that almost no pilot can meaningfully use.
The current FAA budget contains a myriad of objectives. Integration of Unmanned Air Systems (UAS) into the national aerospace system is one. Advocacy of U.S.-based companies is another. Flight safety is in there somewhere as well. Across the pond, the European Union Aviation Safety Agency (EASA) has targeted 2027 as an entry point for single-pilot air carrier operations (initially, it will be limited to long-distance cruise flights, with two pilots still required in the terminal area). Boeing President Alexander Feldman commented on the issue in November. “The technology is there for single pilots; it’s really about where the regulators and the general public feel comfortable.” Recently, Boeing CEO Dave Calhoun echoed the sentiment, stating, “The future of autonomy is real.” Airbus has long advocated for the removal of the second pilot. Tony Lucas, President of the Australian & International Pilots Association, counters, “The people going down this route aren’t the people who fly jets every day. When things go awry [in an aircraft], they go awry fairly quickly.”
Technology vs. Humanity
The days of pilots having direct control over their aircraft are rapidly diminishing. Computer nannies have been introduced into nearly every element of flight. In most modern jets, basic control inputs are “modified” by computers. It used to be that pilots pushed the stick, and that is where the aircraft went. It was reliable but not always safe (humans occasionally overreact). Limiting control surface deflection to maintain safe airframe loads (or to prevent stalls) has undoubtedly saved lives, but it has also produced an uncomfortable byproduct. In a fly-by-wire aircraft, pilots now issue maneuvering “requests” instead of commands. Automation has the final say over an increasing number of processes. This has produced a pretty good safety record, but it is not without danger. Look to Boeing’s own 737 MAX as an example.
Technology has proven to be highly effective in reducing the accident rate in aviation, but up until now, it has always been in conjunction with a human operator. In the 1950s and 60s (the dawn of jet travel), there were half-a-dozen fatal crashes every year. Now they are exceedingly uncommon. And when they do occur, they rarely result in mass casualties (the last in the U.S. was the crash of Colgan Air 3407 in 2009). The proliferation of reliable turbine engines and a better understanding of adverse weather is a large part of the massive improvement in air safety over the past 60 years. Technological innovation has done the rest, yielding a 95 percent increase in airline safety since 1998.
General aviation is a slightly different story. Though the accident rate has decreased, it has not been as quick. There are a few obvious reasons for this. The first is that GA is comprised of a considerable proportion of piston aircraft. The reliability of piston powerplants has not changed much over the past half-century. The second is that the GA fleet has the tendency to be older on average (GA aircraft log substantially fewer hours per year than airliners, giving them a longer lifespan). Also, GA does not benefit as much from risk analysis. Dedicated safety departments at airlines are staffed by full-time professionals who tap into data from thousands of flights per day. This data provides constructive feedback into training programs. Part 91 remains something of the “Wild West” in comparison. Still, there is hope on the horizon. ADS-B data provides more information for accident investigators who research small aircraft crashes. Better data will hopefully drive emerging training standards and lead to new technologies.
Automation is a big contributor to safety, but up until now, it has occurred mostly in bite-sized pieces. As a case example, GPWS was developed to address a specific threat (controlled flight into terrain). Later it was “enhanced” with several additional modes that improved pilot responses. Synthetic vision takes the EGPWS solution and presents it more intuitively. Similarly, TCAS originally alerted pilots to proximate aircraft. Now it provides unambiguous directions to avoid a collision. Some aircraft will automatically fly avoidance maneuvers through the autopilot. If you want to fly safer, technology has proven a highly efficient solution.
Single Pilot Operations
China Eastern Airlines recently partnered in an academic paper about Single Pilot Operations
(SPO) in conjunction with Shanghai Jiao Tong University. Much attention throughout the paper is
focused on the financial aspect of SPO. Likewise, a NASA Technological Interchange Meeting (TIM) focused on the issue in 2012. Again, most participants identified the primary driver of SPO to be the economic benefits to airlines. However, there was some disagreement as to the scale of savings (or whether it would be cheaper once the costs of new infrastructure were included).
The Union Bank of Switzerland (UBS) predicts that cost savings to the airlines of SPO would amount to $15B per year worldwide. This analysis most likely focuses on the direct costs of first officers (a role that would be modified by SPO). More difficult to analyze are the ancillary costs required to implement SPO on a broad scale. For example, a low latency, highly secure air-ground link would be required in the event a pilot became incapacitated or erratic. New air traffic control procedures would have to be incorporated as well. Multiply this on a global scale and the challenges to infrastructure are clear.
Then there is the question of equivalent safety. Accidents per 100,000 hours are 42 times higher in GA compared to the airlines. Corporate operators fare better, yet are still 10 times more likely to have an accident than their airline peers. SPO is undoubtedly a contributor to this comparative mismatch. Loss of Control In-flight (LOC-I) is a particularly deadly form of aviation accidents. Statistics have demonstrated that single-pilot operations produce 6.7 times more fatal accidents than multi-crew operations due to loss of control.
The movement towards SPO will undoubtedly enhance flight safety for GA, regardless of what it does to the airlines. Yet the transformation may not be particularly quick. Optimism from manufacturers revolves around the generational pace of technological innovation. In 1950, for example, many aircraft required five crewmembers (a radio operator, navigator, flight engineer, and two pilots). The reduction from five to two took two decades. We have required that number for 40 years now. A mad scramble is underway to unravel this. At stake is supremacy in the manufacturing battle for the next generation of aircraft. The FAA, in part, is commissioned to facilitate the ability of U.S.-based aviation companies to compete with international conglomerates. The EASA does the same for European plane builders. The first manufacturer to produce an SPO-certificated aircraft (with an equivalent level of current safety) will possess a massive advantage.
The Displacement of Skilled Workers
The question falls into many parts. Is technology better for safety than a second pilot? If we can replace an airline pilot with automation, is GA on the brink of pilotless aircraft? What will SPO infrastructure cost to develop, deploy, maintain and operate? Are we rapidly approaching a new horizon in which automation kills the need for skilled workers? If technology can replace a pilot, how far behind is medicine, computer programming or law?
Just when concern begins to peak, another technological glitch saves the day. No matter how well our devices are engineered, there is always room for failure (or antagonistic exploitation). Automation solves very narrow problems and solves them very well. The complex stuff has traditionally been left to human operators. There is a good reason for this. Computers are great at quickly and reliably solving problems that exist within definite boundaries. Once the probabilistic haze of the real world is introduced, the potential for fatal errors becomes very real. The first generation of SPO airliners will undoubtedly come with hard lessons. It is easy to point to the fact that pilot error is the leading cause of accidents, but this glosses over a crucial element: the ratio between accidents caused and accidents prevented. A crash is obvious and comes with reporting requirements. A pilot saving the day is reminisced over a beer. Are humans worth the price? Will autonomous flight produce a new frontier, or will it drift along like the decades-old promise of a flying car? Either way, we are bound to discover the value of a pilot eventually.