Menu

See & Avoid Doesn’t Work

Contemplate the worst scenario that might confront a pilot during a flight. What comes to mind? Fire? Flight control failure? Engine failure? Perhaps it’s flight crew incapacitation, explosive decompression or severe structural damage.

No doubt about it, those all fall into the Very Bad Day category. But there’s one that can be even worse: a mid-air collision. That’s because it can involve all the problems listed above — at the same time. And since the parties involved aren’t aware of the impending crunch until it’s too late, the mid-air is usually accompanied by a violent element of surprise, confusion, and initial denial.

You might think fatal mid-airs are rare events, and from a purely statistical standpoint I’d have to agree. According to the 2010 Nall Report, a fatal mid-air occurs about once every 8 million flight hours. Think of it as the roughly the same odds as winning the lottery or being struck by lighting. Doesn’t sound so bad, does it? A typical GA pilot might accumulate but thousand or so hours over a full lifetime of flying.

So what’s there to worry about? Plenty. The “big sky” theory may sound good, but it doesn’t hold up very well under close scrutiny. It’s true that the navigable atmosphere over the United States alone is massive — about 20 million cubic miles — and there are relatively few airplanes in the sky. Even on those occasions where a collision is possible, modern tools such as radar, TCAS, VHF communication, and anywhere between two and four sets of eyeballs almost always succeed in averting the disaster. If aircraft were equally distributed throughout the atmosphere, the “big sky” idea would be pretty comforting.

But airplanes cluster near airports, large cities, and on thin slices of the sky known as “airways”. For the VFR types, airspace and terrain often crowd planes into small swaths of the air in places like the Santa Ana Canyon or Banning Pass. The sky is much like the ground: vehicles stick to relatively confined spaces and that makes collisions a serious hazard.

Since we’re on the topic of statistics, let me give you a few of my own: I personally know two people who have been struck by lightning, and a winning lottery ticket was recently sold not 300 feet from my front door. Hey, crazy stuff happens. But unlike lighting strikes and golden tickets, we’re not all facing the same odds. The risk profile varies widely depending on the type of flying you’re doing.

For example, flight instruction is frequently a factor; thirty-seven percent of mid-airs occur with a CFI on board. Many instructional flights happen near airports, and as previously mentioned, that’s where other airplanes tend to congregate. On the other hand, if you fly airliners, your risk of a mid-air is rather low because the aircraft itself is large and easy to see, you’re always flying IFR, and the most sophisticated traffic avoidance hardware available is always installed. Airliners also spend most of their time in cruise and are in constant radar contact with ATC.

Midair collisions are almost as old as powered flight itself.  This B-17 collided with a German fighter over Tunisia in 1943.

Midair collisions are almost as old as powered flight itself. This B-17 collided with a German fighter over Tunisia in 1943.

Think it can’t happen to you? Think again. Some very talented, capable, and well-respected pilots have been involved in mid-air collisions. I know a guy who was involved in one while flying a large-cabin, TCAS-equipped business jet under Instrument Flight Rules. Alan Klapmeier, the founder of Cirrus Aircraft, was in one too. Richard Collins, famed Flying columnist, was in a mid-air. Speaking of Flying, the recent Editor-in-Chief owns a very nice Cirrus SR-22 which was in a mid-air. And lastly, a decade ago I was in a mid-air collision myself.

I’ll save the blow-by-blow (no pun intended) on that for another day. The point I’m trying to make is that the odds of a mid-air are probably greater than you think, especially if you live in a populated metropolitan area and fly VFR. If you’ve ever had a close encounter with another airplane in flight, you were only separated from “those who have” by nothing more than a miniscule sliver of plain old luck.

Think about that for a moment.

This may be hard to believe, but there is some good news. For one thing, mid-airs are not always fatal. It seems intuitive that most collisions would involve fatalities, but all the people I cited above survived, including (obviously) myself. Also, technology is rapidly advancing, from cheap TCAD boxes to airframe parachutes to super-bright LED exterior lighting.

The question we should all be asking ourselves is how we avoid ending up in a mid-air, fatal or otherwise. If you refer to official guidance from the FAA, the answer is to simply look out the window and spot the other airplane before it hits you. This technique, referred to as “see and avoid”, is still considered adequate for preventing collisions. Here are a couple of passages from Chapter 1 of the Airplane Flying Handbook:

The “See and Avoid” concept relies on knowledge of the limitations of the human eye, and the use of proper visual scanning techniques to help compensate for these limitations. The importance of, and the proper techniques for, visual scanning should be taught to a student pilot at the very beginning of flight training.

Proper clearing procedures, combined with proper visual scanning techniques, are the most
effective strategy for collision avoidance.

Other FAA publications, ranging from the Aeronautical Information Manual, to Advisory Circulars like AC-90-48 (“Pilot’s Role in Collision Avoidance”) will give you the same spiel: “see and avoid will keep you safe”. And it will! Until it doesn’t.

From my perspective as someone who’s been in a mid-air and who was using proper clearing and scanning techniques at the time, I take it as gospel that “see & avoid” won’t always do the trick. I’m just one guy, of course. But many others — some institutional in nature — just happen to agree with me.

For example, a couple of years ago Canada’s Transportation Safety Board issued an accident report on a mid-air collision between a Beech V-35B Bonanza and a PA-28 Cherokee over northern Virginia. Canada was tasked with performing the investigation because the pilots of the Bonanza were employees of the NTSB while the Cherokee was piloted by an employee of the FAA.

I won’t keep you in suspense. The conclusion from the TSB was that the “see and avoid” concept was inadequate. They even quoted a 1991 report produced by the Australian Transport Safety Bureau which provides an overview of the major factors that limit the effectiveness of the see-and-avoid principle in preventing mid-air collisions, as well as a 2005 scientific study published in Aviation, Space, and Environmental Medicine which came to the same conclusions.

The main points:

  • Cockpit workload and other factors reduce the time that pilots spend in traffic scans, and even when pilots are looking out, there is no guarantee that other aircraft will be sighted.
  • Visual scanning involves moving the eyes in order to bring successive areas of the visual field onto the small area of sharp vision in the center of the eye. The process is frequently unsystematic and may leave large areas of the field of view unsearched.
  • A thorough, systematic search is not a solution as in most cases it would take an impractical amount of time.
  • The physical limitations of the human eye are such that even the most careful search does not guarantee that traffic will be sighted.
  • The pilot’s functional visual field contracts under conditions of stress or increased workload. The resulting ‘tunnel vision’ reduces the chance that an approaching aircraft will be seen in peripheral vision.
  • The human visual system is better at detecting moving targets than stationary targets, yet in most cases, an aircraft on a collision course appears as a stationary target in the pilot’s visual field.
  • An approaching aircraft, in many cases, presents a very small visual angle until a short time before impact.
  • Complex backgrounds such as ground features or clouds hamper the identification of aircraft via a visual effect known as ‘contour interaction’. This occurs when background contours interact with the form of the aircraft, producing a less distinct image.
  • Even when an approaching aircraft has been sighted, there is no guarantee that evasive action will be successful.
  • Because of its many limitations, the see-and-avoid concept should not be expected to fulfill a significant role in future air traffic systems.
  • Transportation Safety Board of Canada aviation investigation report A06O0206 identified that there is a high risk of mid-air collisions in congested airspace when aircraft are not alerted to the presence of other aircraft and rely solely on the see‑and-avoid principle.

There’s one more area of the TSB report which is worth of quotation. In it, they reference a British Royal Air Force study into mid-air collisions. If you’re keeping score, that’s the third sovereign agency to reach the conclusion that “see and avoid” is inadequate. Yet our own FAA, which oversees about 80% of the world’s aircraft and almost all of the high traffic density airspace, still officially proclaims that one can look out the window and see everything that needs to be seen.

This accident has demonstrated yet again that relying solely on the see-and-avoid principle to avoid collisions between aircraft operating under visual flight rules (VFR) in congested airspace is inadequate.

A number of international studies have addressed the overall issue of the effectiveness of the see-and-avoid principle, as well as the risks of collision associated with this principle. All acknowledged the underlying physiological limitations at play and that, when mid-air collisions occur, “failure to see-and-avoid is due almost entirely to the failure to see.”

One study stated that “our data suggest that the relatively low (though unacceptable) rate of mid-air collisions in general aviation aircraft not equipped with TCAS [traffic alert and collision avoidance system] is as much a function of the ‘big sky’ as it is of effective visual scanning.”

A British Royal Air Force study into mid-air collisions, which were deemed to be random, found that the probability of conflict is proportional to the square of the traffic density, and recommended avoiding altitude restrictions that concentrate traffic.

Measures such as improving aircraft conspicuity, pilot scanning techniques, and pilot traffic awareness can reduce risks, but they do not overcome the underlying physiological limitations that create the residual risk associated with a see-and-avoid method.

It’s obvious that “see and avoid” cannot, by itself, ensure our safety. If it could, there’d be no need for TCAS or most of our controlled airspace (both of which came about because of high-profile mid-air collisions, I might add!). I’m not necessarily in favor of mandating any additional equipment, airspace, or restrictions, especially on general aviation. But it’s clear that serious changes are needed in how collision avoidance is taught, especially as it concerns “see and avoid”. The concept has serious limitations which must be understood so the pilot-in-command can make educated decisions about how — or even if — they want to mitigate those risks.

I sincerely hope our nation’s regulatory and safety organizations will eventually acknowledge what we all know to be true: “see and avoid”, while a good start and certainly a vital part of collision avoidance, is simply not sufficient to ensure traffic separation.

Ron Rapp is a Southern California-based charter pilot, aerobatic CFI, and aircraft owner whose 9,000+ hours have encompassed everything from homebuilts to business jets. He’s written mile-long messages in the air as a Skytyper, crop-dusted with ex-military King Airs, flown across oceans in a Gulfstream IV, and tumbled through the air in his Pitts S-2B. Visit Ron’s website.

26 Comments

  1. Buckholz Traffic

    November 12, 2015 at 7:01 pm

    I didn’t find it in the list, but the mere physical configuration of airplanes makes about 1/3 of the sky unseeable. I can’t see below my low wing airplane and I can’t turn my head far enough to see most of what’s behind me. It’s a probability game and without some form of automated warning its just a matter of time before a mid-air occurs, no matter how vigilant the pilots are.

    • Very true. In my midair, I was flying a biplane, which has the disadvantage of being built with wings both above and below the pilot. After the incident (in which nobody was hurt, thankfully), I obtained diagrams of the aircraft and analyzed the myriad ways a pilot’s visibility was limited. It was pretty eye-opening. Even if your head is outside of the cockpit 100% of the time, there are huge chunks of sky you’ll never see even if you S-turn. And let’s face it, who does that in cruise, during descent, or — to be honest — most phases of flight?

      Like I said in the post, not everyone faces the same risk. Pilots engaged in local VFR flying — probably the majority of AOPA’s membership — are at greatest risk because they spend time near airports and other places where aircraft tend to congregate.

  2. I am a student pilot. Two weeks ago I was flying patterns and touch and go’s around my local airport following a right downwind pattern. As I was about to leave my right downwind leg to base and final approach, cleared by the tower, the tower radioed me to watch out for a business jet at my one o’clock…It was entering base from a left airport approach and had overshot its base to final turn. Simultaneously, I spotted the jet, much to close for comfort, about to cross my flight path just above me. I reacted before my instructor, immediately pushing the yoke down and right, as the jet crossed above me, made a somewhat shaken final turn into a full landing, and then tried to find out what the heck had just happened!

    • Good work! And good lesson, too: controllers make mistakes. Even when they don’t, in a busy airport environment, even if they’ve got the flick (i.e. situational awareness), they can’t look everywhere at once and might not see that a pilot isn’t doing what he or she was instructed to do. Who knows, there could even be a reasonable explanation for the overshoot; maybe the jet overshot that turn because they had to go around a flock of geese or were avoiding another aircraft themselves.

      I wish I could say you’ll never encounter another aircraft at close range again, but the truth is that you probably will. Keep those eyes out, use all your resources, and fly smart…. oh, and most of all — have fun flying!

  3. Ron,
    How often have we been alerted by ATC of nearby traffic that we never get a visual on ? more times than I like to admit and that’s AFTER I was warned about such traffic. So yes, I agree that see and avoid is very limited in it’s effectiveness but most mid-air collisions happen around non towered airports as far as I know and again, I have to come back to my comments from a few weeks ago about pilots flying around with no respect to procedures and with no radio’s, no position reports etc..
    I’m amazed there are such few accidents after all. What are these guys going to do after 2020 when the ADS-B becomes mandatory ??? If they can’t afford to install a cheap radio how are they going to manage the cost of ADS-B is my question. I believe that the FAA should make radios or some form of communication mandatory on ALL aircraft that share our airspace regardless of what class of aircraft it may be. And not only that, I would impose stiff penalties for not following the proper procedures.
    At least, that would be a start and at the very least, avoid some mid-airs.
    Rene

    • If they can’t afford a cheap radio, they’re not going where ADS-B will be required.

      • I would be fascinated to see an airplane with ADS-B Out but no communications radio. I bet it will eventually happen, but probably because a NORDO situation rather than a lack of equipment in the plane. 🙂

    • Agreed! It’s amazing how often we don’t spot traffic and/or how long it takes to find it — even when it’s a large airliner, and I’ve got ATC pointing it out AND a TCAS screen showing the exact location. Hazy windows, harsh light or glare, and a big sky to scan are substantial barriers to spotting aircraft, even if you know exactly where they are.

      Aircraft generally have to have two way radios in B, C, and D airspace, which is where most of the large congested airports are located. The problem with mandating them for all aircraft is that powered parachutes, weight shift airplanes, ultralights, antiques, gliders, etc are not terribly well suited for radio communication even if you have a radio. Open cockpit biplanes are another such example. Even with a good headset and a good radio, the ambient noise is so high that you can barely make effective communication work.

      I do agree with you on proper procedures in the pattern and such… but that’s hard to prove after the fact. I think we have to police ourselves rather than asking the FAA to do it. I doubt they have the resources to accomplish that kind of thing at the nation’s thousands of non-towered fields anyway.

  4. The Australian study also pointed out the ineffectiveness of lights during the day time, yet the FAA mandates that the anti-collision lights be on and recommends that landing lights be on. During the daytime, against the sky as a background (aka the horizon), the eye detects the contrast with the light that the aircraft generates. Turning on the landing light reduces the contrast and makes the traffic more difficult to see. During WWII, the Yehudi project used lights on the leading edge of a B24 to make it less visible to submarines on the surface, IOW turning the lights on is a stealth device. See https://en.wikipedia.org/wiki/Yehudi_lights

    • What an interesting subject, John! Thanks for pointing that out. It’s counterintuitive to think that turning on lights would make something LESS visible, but after reading about the Yehudi lights, I can see how that would be possible. I’m sure it depends on the aircraft, the type and location of lights, and the sky condition.

      • It definitely depends on the sky condition and the background. Next time you are sitting on the taxiway waiting to takeoff during a clear day, watch the aircraft approaching the runway and observe when you can determine if they have a landing light on. It will surprise you. You will pickup the approaching aircraft several miles away as a dark small image and it will have to be almost on top of you before you can see the landing light.

        What is effective, is a pulsing landing light, because the eye’s rods can detect the contrast change from quite some distance, but a steady landing light is of no use in these conditions and a strobe can’t be seen either.

        Traffic from below and climbing will present a dark background and the landing light helps. Descending traffic or level traffic, the landing light does not help or can even reduce the contrast with a bright sky. Remember the earth is curved and falls away from your altitude at the horizon, so traffic at your altitude has a bright background.

  5. disqus_B1vk25qxNZ

    November 12, 2015 at 9:12 pm

    We glider folks are taking up PowerFLARM which shows Flarm, ADS-B and Mode C/S traffic. Unfortunately transponder targets have no azimuth information and uptake in North America is not as thorough as in Europe where Flarm is widely used in GA.

    The ADS-B targets I see so far are heavies and it’s great knowing their course, altitude and descent/climb rate ten miles out. Gliders come up a couple miles away and you can’t always see them even when you know exactly where to look. So I just make very sure to be well clear until we’re both visual.

    PowerFLARM is $1500 plus $240 for external antenna if your aircraft is metal.

    • Another great resource. Thanks for pointing that out! When I learned to fly gliders, we didn’t have radios or transponders, and while we were training in the L.A. basin, it was on the eastern end where things were a little quieter. Of course, we did mix it up with the large jets flying the pattern at March AFB, but they were usually easy to see and relatively predictable.

  6. I’m actually more concerned about the drive to the airport than a mid air. Since general aviation is down some 30% over the last four decades adding more automation and regulation is just another crutch to lean on to make flight 100% safe.

    • Statistically you’re probably right. Of course, the midair is almost certainly going to have a higher fatality rate than an average car crash. The problem with this kind of thinking is that the risk of a midair is not the same in every place. Patterns, airports, practice areas — any place where a lot of airplane congregate — will typically have a higher risk for a midair collision.

      GA traffic is definitely down over the past few decades, but midairs still happen. A busy airport on a weekend can be as much of a zoo today as it was back then. Flying will never be 100% safe, but as they say, knowledge is power. A thoughtful approach to collision avoidance can save a life, and that life might be yours.

  7. You know Ron, People ask me what Im scared about when I’m flying. Engine Failures? Nope. Control failures? Nope. But the possibility of a mid air scares the you know what out of me. When I was scratching my head when my CFI told me “look for traffic” I realized the downfalls when I as at towered airports and the TWR pointed out traffic I never did see.

    I think as long as we know how and where to look it reduces our risk, but even at Cessna speed, opposing traffic. By the time you see someone, you are already up the creek. Closure rate is a matter of seconds and at Jet speeds it can be even less.

    Great article Sir. I’m trying to figure the solution as we speak.

    • I think the best solution is to use all available resources, and use them wisely. Take “looking for traffic”, for instance. One thing Michael Church said that I’ve always repeated is that it’s not enough to glance in a specific direction. Clearing before, say, an aerobatic maneuver means KNOWING there’s nobody there. Not thinking, guessing, or hoping. As a result, as they years go by I spend a lot more time clearing than I ever did before.

      If it makes you feel any better, lower GA activity and increased use of TCAD, ADS-B, and other technology is probably making the odds of a midair lower than in the past.

      • He told me the same thing. He asked me if I was downright SURE there was no traffic before doing that Split – S

  8. Around here, BZN, MT, pilots use the radio instead of their eyes…not good. When I see a plane close enough to be a potential hazard, I waggle my wings. In 1700 hrs of flying, mostly here, I’ve gotten less than a dozen wing-wags back. The radio is not a solution, sorry.

    The radio is a crutch, and a poor one. Think you are not using it thusly? Then I challenge you to turn off the radio, if legal, and then fly for a while with see and avoid only. If you don’t get paranoid, then either you have good s&a skills, or you are clueless :-). In WW1, the pilots had a saying “spot the guy who’s trying to kill you.” I say that phrase to myself often when I fly (usually biplanes, Ron); it really helps sharpen my attitude and improves my s&a skills, I think.

    In my 1700 hrs (all GA), I’ve had 3 situations where I (or my instructor) had to swerve to miss someone…. but no midairs. If you practice s&a, it works. Use your passengers to help! The best passenger I ever had was a young Marine instructor pilot – now that guy could Really spot aircraft…it did my heart good to know that our country was being protected by guys that sharp 🙂

  9. Last week I was to the right of the airport and cleared to land on the short left runway. Another aircraft was to the left of the airport and cleared to land on the long right runway. I was way out in front of the other aircraft and sped up to 80kts. Tower suddenly tells me to turn now behind the airplane, as our paths were converging. I was far ahead of the airplane and cleared to land first, so how was I surprised by the airplane streaking in front of me and I had to take evasive action?
    I was in a helicopter, and that’s why aircraft of vastly different performance abilities need to fly at different altitudes to avoid conflicts. That usually works fine, but sometimes airplanes fly low at heli altitudes (500-700′ agl), and during landings.

  10. See and Avoid is necessary – but not sufficient. My experience matches that of other pilots Italk with. Even when on flight following, and traffic is identified on radar and reported to me, an average of only one out of five are identifiable by See and Avoid. IE: when we are sailing along fat, dumb, and happy, there are obviously many more planes out there than what we think.

    • I absolutely agree, it’s amazing how many airplanes we don’t see, even when we know for a fact we’re looking right at them! Visual scanning is one component of a larger collision avoidance strategy. Considering the consequences of meeting another airplane in the sky, it makes logical sense to avail oneself of every available tool for preventing a collision.

      • Thanks, Ron. When you realize that we DON’T see most of the other planes up there, it’s:
        (A) frightening,
        (B) sobering,
        (C)) enlightening,
        (D) all the above

  11. Ron,
    Two weeks before you posted this article, I was in a head-on mid-air. I was being quite vigilant, and did, in fact see the other plane. But the time between first sight and impact was impossibly short (I had only enough time to gasp) Fortunately nobody got hurt.
    It took over a year for me to get over that jarring experience enough to write about the lesson learned. As I began researching, I discovered that you had already written the article, and did a great job of it. Bravo sir – top shelf.
    Now we must work to erase the institutionalized myth embodied in the “see and avoid” concept of collision avoidance. As you said, most pilots never even see it coming. Although I did, in that head-on, our “slow” 150 knot GA planes closed so fast that avoidance was, quite simply, impossible.
    Thanks for helping to get the word out.
    Cheers,
    Kevin

Comments are closed.