How Airplanes Mostly Stopped Flying Into Terrain And Other Safety Improvements

We have all heard the statistics on how safe air travel is, with more people dying and getting injured on their way to and from the airport than while traveling by airplane. Things weren’t always this way, of course. Throughout the early days of commercial air travel and well into the 1980s there were many crashes that served as harsh lessons on basic air safety. The most tragic ones are probably those with a human cause, whether it was due to improper maintenance or pilot error, as we generally assume that we have a human element in the chain of events explicitly to prevent tragedies like these.

Among the worst pilot errors we find the phenomenon of controlled flight into terrain (CFIT), which usually sees the pilot losing track of his bearings due to a variety of reasons before a usually high-speed and fatal crash. When it comes to keeping airplanes off the ground until they’re at their destination, here ground proximity warning systems (GPWS) and successors have added a layer of safety, along with stall warnings and other automatic warning signals provided by the avionics.

With the recent passing of C. Donald Bateman – who has been credited with designing the GPWS – it seems like a good time to appreciate the technology that makes flying into the relatively safe experience that it is today.

The Art Of Missing The Ground

As Douglas Adams once put it: “The knack [of flying] lies in learning how to throw yourself at the ground and miss”. As quaint as this may sound, it covers the two most essential elements of flying: keeping up sufficient velocity and navigating in a way that keeps one from intersecting with immovable elements such as mountain ranges, buildings and even just flat ground that is not a landing strip. Ideally, an airplane will thus take off, fly a set course and land again at its destination airport. Unfortunately, there are many ways in which this can go (catastrophically) wrong.

Airplane accidents cover a wide range of causes, ranging from weather-related events like downbursts – which can cause instant loss of lift during landing – to mechanical and similar issues. For the latter category famous cases include dodgy wiring causing a mid-air explosion (TWA800), and a lack of lubrication leading to in-flight failure of the horizontal stabilizer (Alaska Airlines 261). Here improved maintenance oversight has led to improvements, but it shares a similar element as the other category: pilot error, which itself is a large amalgamation of factors. Rarely is human error deliberate, but factors like fatigue, distractions, confusion, disorientation and more can all play a role in a disastrous outcome.

Perhaps most tragic here is spatial disorientation, where usually in poor visual conditions the pilot is unable to ascertain what the orientation of the airplane is. In numerous cases, this has led to the pilots giving incorrect inputs on the controls, leading to a loss of lift, attitude and ultimately resulting in a deathly spiral or the airplane simply falling out of the sky due to an aerodynamic stall. A good example of the latter is Air France 447, which occurred after the pilots were handed back control from the autopilot when due to icing conditions affecting the pitot tubes inconsistent airspeeds were registered.

Within minutes, the airplane’s crew had taken a perfectly functioning airplane from stable flight to wild curves and a steep climb, before a complete stall condition caused the airplane to plummet into the waters of the Atlantic Ocean. Despite multiple stall warnings during these wild maneuvers and clear indications on the backup (analog) instruments, the night time conditions with no clearly visible horizon likely contributed to this tragic and uncontrolled plummeting into terrain.

For such accidents, better training, better oversight of maintenance and repair work and adherence to checklists have shown significant improvements. Meanwhile the risk of microbursts has lessened with a better understanding of when they occur and how to react to them. Yet what about the risk of controlled flights into terrain?

Terrain, Pull Up

N2969G, the aircraft involved in the Alaska Airlines 1866 accident, seen at San Francisco International Airport in 1967, while still operating with Pacific Air Lines. (Credit: Jon Proctor)
N2969G, the aircraft involved in the Alaska Airlines 1866 accident, seen at San Francisco International Airport in 1967, while still operating with Pacific Air Lines. (Credit: Jon Proctor)

It was the Alaska Airlines 1866 crash that inspired Bateman to work on a solution for the CFIT phenomenon. This particular flight crashed in 1971 after flawed navigation led to the crew descending too early during its approach, causing it to impact a mountain. Such a controlled flight towards a certain demise had been frustratingly common ever since the dawn of commercial aviation, with the 1936 Havørn Accident in Norway involving a Junkers Ju 52 being among the first recorded incidents.

By the time of the Alaska Airlines 1866 accident, the use of cockpit voice recorders (CVR) and flight data recorders (FDR) was fortunately standard, which gave a much better idea of what the crew saw in terms of instrument data and what their input was to the aircraft’s engines and control surfaces, as well as verbal communications in the cockpit. Although faulty navigation information received by the crew on their radio equipment apparently led the crew to believe that they were closer to the airport than they truly were, the crew was blissfully unaware of their imminent doom until it was too late. What if the crew had received warning about the obstacle and their low altitude?

Called Terrain Awareness and Warning System (TAWS) by the FAA, the original GPWS system as developed by Bateman as an engineer at Honeywell used radio waves to track the airplane’s altitude, along with parameters like descent rate, bank angle and others which can potentially endanger the aircraft if exceeding the known safe range. A major limitation of GPWS is that it only considers what is below the aircraft, which is what Enhanced Ground Proximity Warning System (EGPWS) sought to improve upon. Bateman was also involved in EGPWS development at Honeywell during the 1990s.

With EGPWS, the old system is augmented with more sensors to also look ahead of the airplane, combined with GPS and a database with terrain features including airports. This new system was designed to prevent tragedies like the 1997 Korean Air 801 crash that involved the CFIT at night of a Boeing 747-300 in mountainous terrain on Guam. By effectively creating a virtual corridor in which the airplane moves, any deviations can ideally be quickly noticed and reported to the crew, who can then correct the course.


The Lockheed L-1049A Super Constellation N6902C 'Star of the Seine'.
The Lockheed L-1049A Super Constellation N6902C ‘Star of the Seine’.

Not all kinetic events while the airplane is still fully under control of the crew involve the ground, of course. This was painfully illustrated back in 1956 when a Lockheed L-1049A and a Douglas DC-7 collided above the Grand Canyon. This accident cost 128 people their lives when the two airplanes unexpectedly encountered each other while maneuvering around cumulus clouds. The DC-7’s left wing destroyed the Constellation’s tail, followed by both critically damaged airplanes hurtling towards the ground.

This crash led to wide-scale changes to air traffic control (ATC), as well as the realization of a need for better separation of flights that did not rely on visual detection by the pilots. Following this, the 1958 mid-air collision, of another DC-7 (United Airlines 736) with a military F-100 Super Sabre jet fighter further served to underline the need to merge the ATC for military and commercial flights into one system, leading to the formation of the FAA after dissolving the previous Civil Aeronautics Administration (CAA).

The 1956 Grand Canyon collision would also result in the creation of the Traffic Collision Avoidance System (TCAS) which has undergone many iterations over the decades. At its core it uses a transponder to provide bidirectional communication between TCAS-equipped airplanes. This ensures that the avionics of each airplane is aware of surrounding airplanes, with the possibility to warn the pilot of an impending collision, as well as as automatic avoidance on some airplanes. Theoretically this means that each aircraft is provided with a virtual safety bubble that no other airplane can enter without being tracked. Although not perfect, and not every airplane is equipped with TCAS – mostly smaller airplanes – each incident and near-miss despite TCAS has led to further improvements.

Safer But Not Safe

Every form of travel comes with a certain risk, so the real question is not which form of travel is perfectly safe, but rather how one can minimize the risks involved. Here we can clearly see in the statistics that the risks in the air are fairly minimal, while the risks of landings and take-offs keep increasing. With more and more flights starting and landing at airports around the world, landing and take-off slots become very congested, leading to accidents and incidents involving runway incursions. A recent example of this is the 2024 Haneda Airport runway collision that saw an Airbus A350 practically land on top of a De Havilland DHC-8 (Dash 8). Fortunately a disaster approaching the 1977 Tenerife airport disaster was here narrowly avoided, albeit with the loss of life in the DHC-8 aircraft.

For the past years, airports around the world have increasingly been adding more technology to keep track of not only airplanes in the sky, but also those that are taxiing or standing around the airport. Especially on busier airports it seems that this is the next frontier in air safety. Ironically not in the sky, but while still on the purportedly safe ground, which brings once again to mind the saying about flying being safer than traveling on the ground. Thanks to EGPWS, TCAS and other innovations this is now more true than ever.

Featured image: “Grand Canyon Sunset Through a de Havilland DHC-6 Twin Otter Airplane Cockpit” by Nan Palmero.

13 thoughts on “How Airplanes Mostly Stopped Flying Into Terrain And Other Safety Improvements

  1. The 2024 Haneda Runway incursion highlights a big issues.

    Aviation is still for the most part human first.

    Haneda runway stop bars (effectively traffic lights) were out for servicing for over a week.

    The runway incursion system (RIMCAS) was active and detected the incursion and was ignored.

    Rather than make use of messaging systems like ADS-C for in cockpit monitoring systems to work one pilot simply listens to a radio.

    The Dash8 did not have ADS-B.

    It’s gaping swiss cheese hole one after another all centering in well we can make do if the tech isn’t there.

    I know it’s harder than I imagine but the runway can have plane crossing, plane landing, plane taking off and no one planes allowed states perfect for mutex management.

  2. The common thread here is someone lost focus on flying the plane first and fixing what went wrong second , Unless some major system fails or falls off and the plane becomes uncontrollable or disintegrates.

    Cant help you with running into cumulus granitius, but then again i don’t fly in clouds.

  3. Human factors will remain principle failure points. Newer stuff such as FANS will probably increase ATC radar (virtual) range and accuracy, as ADS transponders do, but the problem with humans will persist. There is no solution to the human problem; there are procedural and technological mitigations, but there will never be a solution to human error – mostly because we cannot remove the human from the control loop due to ‘edge’ cases.

    Three times in the previous two years, I have been issued erroneous instructions by controllers. Two of the three instances were during night or IMC. The last instance had legal issues, and if it were not for my on-board video and audio recording, I would have probably had my license suspended. The controller that screwed up had been working six days per week for about two years.

    The other human factor issue is aircrew competence. For ATRs, new regulations will require an on-line log of pilot training records and issues. Currently, when a pilot goes from one carrier to another, the new employer has no knowledge of a pilot’s proficiency and performance issues at the previous employer. Also minimum ATP experience levels were increased several years ago.

    As for general aviation, pilot proficiency is a dog’s breakfast. I have two friends with a ‘sport’ pilot license; and I do not consider either to be safe, and neither are competent aviators. The airplane in the hangar next to mine is flown by a 78 year old person that should not be driving, much less flying a high-performance (Cessna T210) general av bird. For geezers, a third class is good for 24 months, so I have flagged his registration number for the next year in my ADS-B display, and stay away from him when he’s off of the tarmac (how is that for using tech to increase my safety?).

    Talk all you want about aviation tech – the human problems will persist.

    1. Where were you flying that an ATC had been working 6 days per week for 2 years? Here in the UK ATC is very heavily regulated including rules (elegantly named Scheme for the Regulation of Air Traffic Controllers’ Hours (SRATCOH)) about how much work a controller can do and not working through fatigue.

      1. OP is presumably in the US (as far as I know, no other country has a ‘sport pilot’ license or refers to their equivalent as such)

        US federal regulations don’t go nearly as far as SRATCOH- they only mandate a break of 24 hours at least once per 7 consecutive days (14 CFR § 65.47).

        I’ve never heard from anyone who actually *wants* ATC to be regularly working 6-day weeks, but it’s legal, and as the US also has an unfortunate shortage of qualified air traffic controllers (and/or a failure to budget enough to hire them) so it’s distressingly common.

        1. Regan fired them *all* back in the 80s after they went on strike for better pay and safer conditions. It is sometimes said, to this day, that they haven’t got back to the skill level they were before that happened.
          Another thing to blame on that useful idiot!

  4. There is a stand up bit by an air traffic controller, called “what must go up doesn’t necessarily come down” or something like that. He said the purpose of the ATC is to route planes sufficiently close to each other and mountains to justify their own job. haha.

    In seriousness though I gave a talk about critical safety in aviation and medicine and used a lot of case study examples from both fields. One of the cases was the Kobe helicopter crash demonstrating that even very brief spacial disorientation can be deadly. I cited a military study that entry into instrument meteorological conditions and failure to respond immediately was responsible for something like 25% of training crashes.

  5. I was on the software development team for the GPWS for the B787 program. It was a Pentium x86 with a ton of flash. Updates to the DB also included man-made objects. There exist GPWS versions for helicopters which include things like power lines. Pictured here is an earlier GPWS line-replaceable unit (LRU) which went onto B747s.. The GPWS I worked on lived together with the TCAS (traffic-avoidance), and WX (weather) in one combined LRU

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.