Are All Drone Near-Misses Really Drones?
Whilst there exists a lot of media hype around the subject of drones almost colliding with planes the actual statistics show this is a relatively rare occurrence.
In 2016 there were 70 Airprox reports submitted to the Civil Aviation Authority (CAA) regarding drone near misses. This compares to 1,258 incidents reported where lasers were shone at aircraft and 1,835 reports of confirmed bird strikes. Up to 31st July 2017 there were 62 drone related Airprox reports submitted.
Undoubtedly the number of Airprox reports involving suspected drones is increasing, but how accurate are these reports? How certain can we be that drones were involved rather than say a carrier bag floating on a thermal?
That may seem a ludicrous suggestion, and no doubt this post will cause condemnation from the professional aviation community, but it’s a very real issue that needs addressing.
One point that illustrates this is that every year prior to 2017 included Airprox reports where pilots identified the objects as model aircraft, yet there hasn’t been a single Airprox report this year involving model aircraft. Are model aircraft pilots being especially cautious or are airline pilots routinely classifying incidents with small objects as drones?
When is a Drone not a Drone?
On April 17th 2016 the media was set alight by reports that a drone had hit a British Airways flight coming in to land at Heathrow Airport in the UK. Widespread condemnation was quick along with increased calls for legislation to tackle this menace. Steve Landells, the by now infamous Flight Safety Specialist for the British Air Line Pilots Association (BALPA) was quick to state it had only been a matter of time before such a collision occurred.
A subsequent investigation however revealed that the flight hadn’t collided with a drone, but by that time the media had moved on and there was little fanfare. The Telegraph reported that rather than being a drone it could have been a plastic bag that had hit the aircraft. Mr Landells didn’t recant his statement and over a year later there has still never been a single confirmed drone collision reported to the CAA.
It seems therefore that pilots can jump to the conclusion an object is a drone simply out of familiarity with the topic. BALPA are in full-time campaigning mode against drone use that they claim endangers aircraft and regularly lobby politicians to tighten legislation. Talk to any commercial pilot about drones and they all believe they are being flown regularly around airports.
Most Airprox reports about drones from commercial pilots relate to sightings on approach to land or climbing out after take-off. The reason for this is quite obvious, at these stages in a flight the aircraft is at a lower altitude (below 5,000ft) and flying at a lower speed than normal cruising speed. There is a higher probability of seeing other smaller objects in the air at theses levels such as balloons, bags, drones, model aircraft, and the catchall term Foreign Object Debris (FOD) for any item they can’t categorise.
When pilots end up flying in the vicinity of such objects they file an Airprox report. The UK Airprox Board define it as:
“An Airprox is a situation in which, in the opinion of a pilot or air traffic services personnel, the distance between aircraft as well as their relative positions and speed have been such that the safety of the aircraft involved may have been compromised.”
The Airprox Board then assess the report and publish their findings. What’s important to understand though is that the pilot states whether the object was a drone or not. Drone near misses may well be overstated if pilots incorrectly identify objects.
If this is the case we may well be experiencing an increase in reported drone near misses not because more drones are being flown in the vicinity of airports, but because pilots are becoming pre-disposed to identify unknown objects as drones.
A320 Near Miss
Scanning through the published findings of the Airprox Board I took one case as an example to illustrate the typical type of report filed as a drone near miss.
Airprox 2017077 occurred on 17th February 2017 at 4:25pm as an Airbus A320 was on approach to Manchester Airport. The summary states:
“THE A320 PILOT reports that the drone was seen during final approach to RW15 at Birmingham, at 6nm. He was flying stabilized on the ILS at 2000ft. The drone was observed slightly right at approximately 300ft below. There was no imminent danger of collision but they reported it to ATC to alert other aircraft.”
Taking a scale drawing of the A320 I plotted lines of sight from the pilot’s position in the cockpit towards the projected height of the drone.
The pilot reported the drone was approximately 300ft below the aircraft. Taking the blue line to intersect with this height difference revealed that the drone would have been approximately 1,800ft away from the pilot when it disappeared from their field of view.
Assuming that the A320 was travelling at 180kts on approach it would have been covering approximately 300ft per second.
If we take 2 seconds as a timeframe to see the object and identify it this would have placed the drone at approximately 2,400ft away from the pilot when it was first seen.
Whilst it is possible for someone with 20/20 vision to discern an object the size of a typical drone (35cm by 35cm) at a distance of 2,400ft away it is also possible to see a carrier bag at such distance.
For example the average resolution of the naked human eye is approximately one arc minute. The smallest resolvable size at 2,400ft away is approximately 21cm, at 1,800ft it’s 16cm. You can see an the size of a typical drone (DJI Phantom) from both 2,400ft and 1,800ft away, but the ability to differentiate it as a drone from another object is a different matter.
For example you may need to see the sweep of propellors to be able to conclude that it’s a drone, or you may need to see the anti-collision strobe lights flashing.
Being 300ft above the drone may rule out the ability to see the lights on it as these are positioned directly underneath the arms of the drone. Also this Airprox summary doesn’t mention that any identifying lights were seen on the drone. The propellors for the DJI Phantom measure around 23cm, so there is a possibility of being able to see the propellors at both 2,400ft and 1,800ft away. However beyond 2,400ft it would be impossible to see the propellors.
Time to Impact
In the above example of the A320 I’ve given 2 seconds as the time from first seeing the object despite the fact that no timeframe was given in the Airprox summary. I took this position to give the pilot the benefit of the doubt and weight the calculations in their favour.
In the 2017 edition of the Airprox Board’s annual magazine, “Airprox”, an article entitled, “When every second counts” gives insights that may indicate this timeframe is overly generous:
“research shows that in normal circumstances the average pilot and aircraft needs anything from nine to 12.5 seconds from spotting another aircraft to processing the closure geometry and avoiding a potential collision.”
“Naturally, the odds of spotting a potential collision reduce in relation to time spent looking out, and the best rule of thumb is 80:20 – 80 percent of the time looking out and just 20 percent inside the cockpit in small chunks.”
“Quite apart from the physiological limitations, the eyes are vulnerable to other visual distractions; lighting, foreign objects, illness, fatigue, emotion, the effect of alcohol, certain medications and age all play their part. Then there are additional challenges such as atmospheric conditions, glare, deterioration of transparencies, aircraft design and cabin temperature, which all take their toll on your eyes and what you can see.”
“It’s a curious thing about flying that many pilots believe they keep a good lookout when in reality it’s less-than-effective; glancing out and scanning with smooth and continuous eye movements is incorrect because for the pilot to perceive another aircraft, time is needed for a stable image of it to fall on the centre of the retina and the pilot’s attention directed towards it.”
If we take the principle that the pilot first saw the object 5 seconds before it was out of their field of view it would have been approximately 1km away from them. Even taking into account the best resolution possible for the human eye it’s stretching the bounds of belief that a drone would be noticeable by a pilot from 1km away as they are on approach to land.
The reason I’m raising this issue is because we seem to have a very one-sided discussion of drones and aircraft in the media at present, which is partly due to BALPA’s lobbying and PR work. Pilot testimony is taken as gospel in a media climate that is still waiting with baited breath for the first drone mid-air collision.
Yet at the same time we see numerous stories relating to airline pilots being under the influence of drugs and alcohol whilst flying, one of which was jailed in March 2017. Writing for the Independent, Simon Calder explained the way that pilots are presently checked to see if they’re under the influence:
“At present, a typical scenario is that ground staff or fellow crew members suspect that a pilot may have been drinking and either challenge him or her or report their fears to the authorities.”
Surprisingly there is no mandatory breathalyser check before taking the controls of the aircraft, however Dr Rob Hunter for BALPA states:
“The data suggests that there is not a problem of drug and alcohol misuse in large commercial air transport operations.”
However, if pilots aren’t routinely breathalysed before flying how can the data be reliable? There is also the issue of fatigue that affects commercial airline pilots. Less than two weeks ago BALPA issued a press release warning of the effects of pilots being pushed to their limits:
“Pilots are proud to be getting a record number of customers off on their holidays this summer but are warning that demanding schedules, lax controls of pilots’ hours of duty and a failure to recruit adequate numbers of pilots, are pushing the system to the limits.”
“BALPA says it is seeing increasing numbers of pilots who are looking to go part time or have become long term sick as a result of fatigue and “burnout” caused by inadequate rest and unworkable patterns of duty.”
In a situation where an airline pilot may be under the influence of alcohol or suffering from fatigue and they notice an object thousands of feet away is it possible that they may jump to the wrong conclusion? Especially when they have only a second or two to see the object and recognise it?
I’m not saying that drones aren’t being flown dangerously or irresponsibly by a small number of individuals, nor do I support such reckless stupidity. What I am saying is that it is a relatively rare occurrence for a drone to be reported as flying in the vicinity of an aircraft. Even then there exists the very real possibility that a significant proportion of them may be FOD misidentified as a drone.
If an airline pilot can misidentify hitting a plastic bag as hitting a drone it’s also possible that sightings of other drones may also be similar debris.
In the years prior to drones being in popular culture there were very few Airprox reports related to drones. In the 5 years from the start of 2010 to the start of 2015 there were only 10 reports in total. Then in 2015 the numbers massively increased and this increase corresponded with widespread reporting of drones in the media.
Prior to 2015 it is likely that when pilots saw Foreign Object Debris such as pieces of plastic floating on thermals they dismissed them as such. Now that pilots have become more acclimatised to the belief that drones are being flown at several thousand feet around airports are they reporting FOD as a drone?
If it was stated that UFO sightings were on the increase would it be correct to surmise that they’re genuine UFOs? As it happens the Independent reported just such an increase in February of this year. Just as we would question whether a pilot had seen a real UFO so we should question whether they’ve seen a real drone.
As the UK Government now moves onto the phase of looking into the likelihood of drones being involved in mid-air collisions isn’t it time that there was some research into the possibility of identifying a drone as a drone? If not then the data everyone seems to be relying upon to show an increase in incidence will undermine the reliability of its findings.