Bird strikes and the light aircraft

The bird strike that everyone remembers ended in the Hudson River with an Airbus and no fatalities. The ones that kill people happen in much smaller aeroplanes, and almost nobody writes about them. Of the 52 deaths recorded in the FAA's National Wildlife Strike Database since 1990, 50 were aboard light aircraft — piston singles and twins, light turboprops and jets, helicopters — across 23 separate accidents. 18 of those accidents happened en route, not in the circuit, and 8 involved a strike on the windscreen.

That is the first thing the data says that the safety leaflets do not: for the aeroplanes most of this audience flies, a bird strike is a structurally different event than it is for a larger one. A strike damages a light aircraft 23.8% of the time. For aircraft above 5.7 tonnes the figure is 6.6%. Broadly overlapping hazards, very different outcomes — about three and a half times the damage rate.

Look down the windscreen row. When a bird hits the windscreen of a light aircraft it damages it 16.6% of the time; on a larger aircraft, 1.4%. This is not a difference in luck. Transport-category windscreens are certified against penetration by a four-pound bird at the design cruising speed at sea level (14 CFR 25.775(b)), and transport-category helicopters carry a 2.2-pound version of the same requirement (29.631). Nothing of the kind applies to a Cessna 172. Its windscreen is there to keep the weather out.

The same asymmetry runs through the whole table. The light-aircraft column is worse in every single row, and on the surfaces that meet the bird head-on — windscreen, nose, radome — it is worse by a factor of roughly five to twelve. Only the propeller, the engines and the tail come close to parity, and that is small comfort: an engine is a worse thing to lose in a single than in a twinjet. A light airframe has less mass to absorb the same impact energy, less redundancy behind each surface, and no engineering margin bought specifically for this.

The season is the part most commonly imported wholesale from European guidance, and it transfers only in part. The EGAST leaflet does name the autumn migration — September and October — so the American peak is not a month European guidance overlooks. What travels badly is the emphasis around it. The leaflet's sharpest warnings fall on March, which it calls the peak of migration intensity, and on June and July, when inexperienced juveniles are learning to fly. Neither is where the American light-aircraft record is heaviest: the peak is September, with August and October close behind, driven by an autumn migration that runs down flyways the length of a continent and takes months to pass. March sits near the quiet end of the American year.

So a European leaflet read in the United States gets the autumn right and the emphasis wrong. The months it marks hardest are not the months this record is heaviest, which is worth knowing before a calendar written for one continent is used to plan caution on another.

Here the data complicates the standard advice rather than confirming it. The leaflet's "plan to fly as high as possible; most birds fly at only a few thousand feet AGL" is true about how often a strike happens: over half of all light-aircraft strikes happen on the ground roll or below 500 feet, and just under two-thirds of those where a height was recorded at all. But the damage rate moves the other way. A strike on the ground damages the aircraft 19% of the time. Above 3,000 feet it is 37.6% — twice as bad.

The honest reading is not that altitude causes damage. It is that altitude selects for the worst encounters: the birds that cruise at 3,000 feet are geese, vultures and raptors, not sparrows, and they are met at cruise speed rather than at rotation. Climbing reduces exposure. It also raises the stakes of the exposure that remains.

Speed is where the standard advice holds up completely. Impact energy goes with the square of the closing speed, and the data follows: below 80 knots a strike damages the aircraft 21.8% of the time, and above 160 knots it is 52.6% — better than even odds. Speed is the variable the damage rate answers to most cleanly, which is what makes it the one figure in this record worth quoting on its own.

Darkness looks dangerous at first glance: strikes at night damage the aircraft 27.9% of the time against 22% by day. But most of that gap is not birds. It is deer — nocturnal, heavy, and standing on the runway; 882 of the night strikes involve a mammal rather than a bird. Count birds alone and most of the gap goes: 23% after dark against 21.1% by day. What is left is small but real, and dusk is worse than either. Birds do fly at night, and the strike record says so — but the animal most likely to ruin a night landing has four legs.

And the warning rarely comes. Only 17.6% of the 30,211 light-aircraft reports record a prior warning of bird activity — though the field is often left blank, and among the reports that answered the question at all it is 30.3%. Either way, most of these pilots were surprised. The tools are thinner here than the folklore suggests on either side: US sectionals do chart wildlife refuges, and AIM 7-5-6 asks for 2,000 feet above them, but nothing charts the seasons or the species the way German aeronautical charts do, and the radar-driven forecasts that do exist are aimed elsewhere: the Air Force's Avian Hazard Advisory System, public but organised around military routes, ranges and airfields, and BirdCast's nightly migration maps, built for ornithology rather than flight planning.

The animal that most reliably wrecks a light aircraft is not a bird. White-tailed deer were struck 999 times and did damage in 87.9% of them — the worst record of anything in the database, and a reminder that the runway at a quiet field is a food source at dusk. Behind the deer come the vultures, the Canada goose and the ducks, all above fifty per cent, then the hawks. At the bottom of the table sit the sparrows, kestrels and starlings — struck constantly, almost never a problem.

This is why "how many birds" is the wrong question and "which birds" is the right one. A flock of starlings across the numbers is a wash and a logbook entry. One Canada goose is an airframe.

What it actually costs

Most bird strikes are not accidents, and the database records what happened next. Light-aircraft strikes produced 5,208 precautionary landings, 1,323 rejected take-offs, 348 injuries, 928,833 aircraft-hours out of service — about 38,701 aircraft-days — and $143.3M in repairs, adjusted for inflation. Up to 22.2% of these strikes changed the flight.

What the numbers point to

None of this is a checklist, and the database is in no position to write one — it records what happened, not what anyone should do about it. But a few things in it are consistent enough to be worth carrying:

• Speed carries the most weight. It is the one piece of standard advice these numbers back without qualification: the damage rate more than doubles between the slowest and the fastest band.
• Height buys fewer encounters, not gentler ones. Climbing changes how often, not how badly.
• Darkness does not behave the way "birds are asleep" suggests — and in this record the reason is mostly four-legged rather than feathered.
• The deer are the outlier. Nothing else in the database damages an aircraft as consistently, and they turn up at unlit fields after dark.
• Damage hides. What ends up grounding an aircraft is routinely found behind a cowling or in the tail, on the ground rather than in the air — which is one reason the thousands of precautionary landings above are hard to read as over-caution.
• The counts are a floor. US reporting is voluntary, so every total here is the strikes somebody chose to file, not the strikes that happened. Reports go to the FAA at wildlife.faa.gov.

Per-airport, per-species and per-aircraft breakdowns of the whole database live in our wildlife strike tracker.