Force of nature

On 26th July 2005, the space shuttle Discovery hit a turkey vulture. The encounter, which happened during liftoff, was caught on video. You see the great bird soaring, perhaps delighting in the awesome thermals of a rocket liftoff, and then abruptly bouncing off the external fuel tank and plummeting, Icarus-like, into the exhaust plume. Because the craft had just begun to accelerate, the damage was small, and largely confined to the vulture. Nonetheless, recalls Travis DeVault, a wildlife biologist with the Sandusky, Ohio, branch of the National Wildlife Research Center, “it created a lot of heartburn.”

The NWRC branch offices share the grounds of Nasa’s Plum Brook Station. Plum Brook is where Nasa tests rocket engines and Mars landers and the like to be sure they’ll perform under the stresses and unusual circumstances of space travel. Plum Brook engineers can create wind that blows six times faster than the speed of sound, and vibrations as ungentle as those of a rocket launch. A wayward buzzard seems laughable by comparison, but no one was laughing. Travis reminds me that it was a ding to space shuttle Columbia’s exterior on the way up that led to the tragic explosion on the way back down.

Plum Brook Station is headquarters to the National Wildlife Strike Database, co-managed by the FAA (Federal Aviation Administration) and the USDA. The FBI is here too, behind a row of closed doors with nameless nameplates, across from Travis’s office. He doesn’t know what they do in there but is impressed by their paper-shredding capabilities. “Comes out like dust.”

In 2015 the National Wildlife Strike Database summarised data from 25 years of collisions between civil aircraft and wildlife. The report goes species by species: number of strikes, the total cost of damage to aircraft, and the number of people killed and injured. There are two ways to be a bird of interest to the FAA (or, who knows, the FBI): weigh a lot or travel in groups. By dint of their size, turkey vultures rank high on the worry list: 18 people injured, one killed, and costly damage to the aircraft 51 percent of the time. By contrast, of 27 documented chickadee strikes, none caused damage of note. To clarify, a turkey vulture is a vulture, not a turkey. Though turkeys, too, have crashed into planes. But only wild ones. Supermarket turkeys have never hit planes, but dozens of supermarket chickens have, because they are fired at jet parts to test their ability to hold up to birds strikes. The device that fires them is called, yes it is, the chicken gun.

The FAA considers the white-tailed deer the most hazardous wildlife to US civil aircraft

Jet aircraft engines are given a “bird ingestion” test, but the birds used in the testing weigh two pounds. Turkey vultures average three. Travis sent me a link to footage of one of these tests, slowed down to reveal a progression of fan-jet engine blades slicing a bird like a meatloaf. With a bird the size of a vulture or a pelican, the blades may end up in pieces, too. Shards of fan blade slamming into delicately calibrated engine parts can have catastrophic consequences. Travis also emailed me control tower footage of a Boeing 757 strike that he uses at trainings for airport biologists. You see an unidentified blackish bird – really just a speck on the screen – disappear into the mouth of an engine during takeoff and produce a near-instantaneous fart of fire out the back. It’s the soundtrack that stays with you. You hear the pilot’s call, “Mayday, Mayday, Mayday,” and in the background the normally cheerful but now darkly sinister twitterings of a bird.

Odds are decent it was a starling – among the six most frequently struck bird species in the United States. To confuse winged predators, starlings sometimes fly in enormous shapeshifting flocks called murmurations that swerve and split and coalesce again, all without warning or seeming logic. The open maw of a jet engine passing through is bound to swallow a few. They’re like avian krill. Worst-case scenario: big birds in groups. It’s no surprise that the species that brought Captain Chesley B. “Sully” Sullenberger down to the Hudson River was Canada geese, one or two per engine.

Travis DeVault’s wildlife strike research has focused, at various times, on turkey vultures, blackbirds, and Canada geese, but tonight he’ll be gathering data on a yet more dangerous creature, the species the FAA considers the “most hazardous wildlife to US civil aircraft”: the white-tailed deer.

From 1990 to 2009, the National Wildlife Strike Database logged 879 collisions between white-tailed deer and aircraft. The impacts injured 26 people and caused, on average, six times as much aircraft damage as other wildlife strikes. Only Canada geese, red-tailed hawks and pelicans caused more harm. Deer strikes have happened on landing, during takeoff, and while taxiing. Obviously no deer have struck planes at cruising altitude. Less obviously, two animals struck planes that were parked.

White-tailed deer are 30 times as heavy as turkey vultures, and they travel in groups. Double whammy, both for aircraft and for vehicles on the road. Travis’s focus, of late, has been this: Creatures on roads and tarmac often don’t get out of the way in time, even when there’s time to get out of the way. He’s looking for answers to the questions we can’t just ask. Why do you just stand there in the headlights? How can we help you? How do you not notice a space shuttle?

Travis is driving me around Plum Brook’s six thousand woodsy acres. It’s a wildlife biologist’s tour of a space centre: See the bald eagle nest up there? And this is a great place to hunt for mushrooms. That building has some big thermal thing. There’s another nest! You can tell which buildings are rarely used, from the deer. A half dozen of them mill about on the lawn of the Hypersonic Tunnel Facility, nibbling and strolling like guests at a wedding reception. Clusters of them appear along the roads with the regularity of mile markers. For the first couple of minutes, I’d lean forward against my shoulder belt and exclaim, “There’s one!” At one point Travis turned to me: “You don’t see many deer where you live, do you, Mary?”

Out where he lives, Travis sees a lot of deer, but he did not see the one he hit. This is often the case. “The deer you hit are the followers,” he says. You brake, and your eye follows the deer you just avoided hitting. “You speed up, and here comes the next one.” As soon as it’s dark, we’ll be meeting up with Travis’s research partner, Tom Seamans, to take some data for what I’m calling the deer in the headlights project.

It’s approaching dusk now, the time of day with the highest hourly incidence of deer-vehicle collisions – three times higher than the dark of night, according to a paper by Travis and four colleagues. Deer are crepuscular, a word born for dermatology but in fact meaning “active at dawn or dusk.” November is the other standout risky drive time, because it’s the rut – mating season. Hell-bent on reproduction, the deer fail to note the most blatant obstacle to the successful onward advancement of their genes: traffic.

“Fast cars have only been around for a hundred years. In terms of evolution, that’s nothing” – Travis DeVault Research scientist

We’re seeing deer all along the roads partly because it’s dusk, and partly because Plum Brook has a lot of them: about ten deer for every one person who works here. Also, a road that runs through woods is an attractive place to a deer. Food grows close by, and it’s a clearing of sorts, so predators can’t easily sneak up unseen. The open space of a road also appeals to birds that hunt insects on the wing, because it’s easier to see and manoeuvre here. The ones that get hit draw the scavengers. Roadkill begets roadkill. (In hopes of preventing another turkey vulture mishap, the Kennedy Space Center ground crews formed a “roadkill posse” to clean up carcasses with unusual alacrity in the days leading up to a launch.) On the simplest level, animals take to the road for the same reason people do: the going is easier.

The low speed limit at Plum Brook has helped keep wildlife populations booming. As long as cars aren’t moving faster than a natural predator would, a prey animal will usually get out of the way in time, even if the driver doesn’t brake. The hunted maintain what’s called a spatial margin of safety. They’re able to visually intuit the distance between themselves and a predator, and they have an uncanny sense of exactly how close they can let that predator come before they need to take off. Flight initiation distance, as that closest point is called, shrinks and lengthens according to circumstances. If animals or birds are feeding on something nutrient-rich and wonderful, they may wait until the last possible moment – the shortest FID – to abandon the feast. If the predator is coming at them at a run, its speed is factored in and they’ll take flight when their pursuer is farther away. They almost always judge the safe getaway distance correctly. Unless the thing coming at them has an engine.

Mammals and birds, sensibly, perceive onrushing cars as predators. Their escape algorithms work well on a congested city street – though you may try, you will almost never hit a pigeon –  but their judgement fails them on a motorway or a rural thoroughfare. Because what predator comes at you at 60 mph? Evolutionarily speaking, this is something new. “Fast cars have only been around for a hundred years,” Travis says. He flips a visor down against the setting sun. “In terms of evolution, that’s nothing.”

Travis has speculated that this explains the baffling inability, among wildlife, to avoid what should be easy to avoid: “a large, noisy vehicle travelling along a predictable path.” Evolution hasn’t had time to upgrade the processors. Judging speed requires an ability to perceive and interpret “looming” – how quickly an object appears to be growing in size as it comes at you. Looming is harder to detect and visually process when the object is travelling quickly. The “looming-sensitive neurons,” as some pigeon researcher went ahead and named them, are overwhelmed.

Travis and Tom have devoted considerable time to studying this. The pair’s original study protocol was straightforward: “We drove a vehicle directly towards turkey vultures …” The vultures had been lured by a raccoon carcass anchored to a heavy metal plate, to keep the birds from dragging it off the road to a more relaxing setting. The vehicle, a Ford F-250 pickup, travelled without braking, at three different, constant speeds: 19, 37, and 56 mph. Flight initiation distance was measured by dropping a beanbag out of the window onto the asphalt to mark the moment a vulture moved to retreat, and then measuring the distance from beanbag to carcass. The FID for 37 mph was not significantly different from the FID at 56 mph, suggesting that, as Travis and Tom had predicted, an unnaturally fast “predator” overtaxes the prey’s sensory and cognitive gifts. No vultures were hit, though there were close calls, all of them at the fastest speed. To see what would happen at even faster speeds, Travis and Tom devised a video truck. Cowbirds – because they’re common around here and hardy – were installed in a roomy cage (and, fret not, released afterward). One wall of the enclosure took the form of a video screen, on which the researchers played footage of a truck driving straight at a video camera they’d placed in the middle of a road. Cowbirds, they found, will take to the air when a vehicle gets to about 100 feet away, regardless of its speed. Up to about 75 mph, they have ample time to get out of the way. At speeds faster than that, the birds made their move too late.

Through the magic of variable speed video playback, Tom and Travis were able to accelerate the video truck to 224 mph –  roughly the speed of a plane taking off. Because that’s really what this research is about: flight safety and aeroplane damage prevention. It would be great to figure out how to prevent the deaths of hundreds of millions of small creatures on US roads each year, but that’s not the ultimate goal of the work. Faced with a truck going as fast as a plane, every one of the cowbirds would have wound up a statistic in the National Wildlife Strike Database.

“When self-driving cars take over the roads, squirrels may no longer be spared by kindly drivers who swerve”

Research on jaywalking pedestrians tells a similar story. Most of our decision-making is based on how far off a car is. We’re not so good at factoring in the speed. Experimental evidence suggests that full looming sensitivity doesn’t develop until adulthood. A young child on the side of the road and a car travelling faster than 20 mph combine to encourage, quoting a team of European psychologists, “injudicious road crossing.” Hence the need for injudiciously punctuated slow children signs. It’s not just that kids aren’t looking when they cross; they’re also not seeing.

For animals facing down a predator, fleeing is but one option. Mammals rely on a diversity of features and behaviours that, over the millennia, have increased their odds of staying alive long enough to pass on their genes. The skunk sprays a vile smell, the porcupine wears darts. When the “predator” is a speeding automobile, these tactics range from ineffective to tragically maladaptive. The turtle stops in its (and your) tracks and pulls its head into its shell. A deer may freeze to avoid being seen among the trees. Squirrels and rabbits zigzag halfway across a street. When your killer is a hawk that’s calculated the likely intersection of its path and yours, changing course abruptly may save your life. When the killer is a land-based commuter, it foils her efforts to avoid hitting you.

When self-driving cars take over the roads, the lives of squirrels and skunks (and cats and small dogs) may no longer be spared by kindly drivers who swerve and brake. By the cold calculus of (human) survival, drivers are safer doing neither. The Centers for Disease Control and Prevention estimates that 10,000 people per year are injured when they take evasive action to avoid hitting an animal. That’s only 2,000 fewer than the number injured when the vehicle actually hits the animal. In 2005, the Insurance Institute for Highway Safety (IIHS) analysed 147 fatal (to the human) vehicle-animal crashes, 77 percent of which involved deer. The initial impact rarely killed or even hurt people; almost always, they died because the driver tried to avoid the animal. The driver or motorcyclist braked, and the vehicle skidded and went off the road or collided with something less yielding than venison.

“The Swedish National Road and Transport Research Institute, which shortly thereafter received a plea to help design a camel crash test dummy.”

The eight exceptions were instances in which large deer – and in one case, a horse – crashed through the windscreen. Taller is a killer. Because now the car’s front end strikes the animal’s legs rather than its torso. And when the legs are knocked out from under, the torso and head pinwheel over the hood and crash down onto the windshield and, if the animal is tall enough, the roof. Thus Volvo has a LADS – large animal detection system – but no SADS. “The camera looks for a specific signature,” a Volvo communications manager said in an email. “A large body mass with four very thin long legs.” The example he gave is a moose. For a 1986 master’s thesis, a team of Swedish bioengineering students staged a moose strike and filmed it at high speed, so the impact and its aftermath could be studied in slow motion. The aim was to provide a more nuanced understanding of the biomechanics of these often devastating collisions, and then use this understanding to develop a moose crash test dummy. An “ill and weak” bull moose “was put to death, and quickly afterwards it was hit by a Volvo 240 at a speed of [50 mph].” The phrasing intrigued me. Apparently the Volvo 240 is a car that goes from 0 to 50 quickly enough to reach a moose in the fleeting moments between its death and the crumpling of its legs. For you could not suspend a euthanized moose in a frame, as this would keep the moose from doing the very thing the authors aimed to study. Anyway. What the film revealed. If the roof collapses as the passenger is thrown forward – and now I borrow the gently evocative phrasing of Swedish moose crash test dummy designer Magnus Gens – “crumbled steel interferes with the head’s path.” Borrowing the less gentle phrasing of “Moose and Other Large Animal Wildlife Vehicle Collisions,” “axial compression . . . causes bony fragments to be pushed into the spinal canal.” Moose falling on driver’s head crushes neck vertebra, sharp pieces of which slice spinal nerves, causing full or partial paralysis. Also disturbingly common: broken face bones and lacerations from hitting the windshield in mid-shatter. The wounds become infected by “debris, hair, entrails, and faeces.” And finally, should the two of you manage to survive the impact, one of you now has a flailing moose in his lap.

Making matters worse: the long legs of a moose may boost the animal’s eyes up above headlight range, eliminating the reflective shine that helps drivers see it in the dark. (The tapetum lucidum, which does the reflecting, is actually there to aid their vision, not ours. It boosts mammals’ vision in low-light conditions by bouncing the light back into the retina a second time.)If you plan to be driving in far northern regions where tall ungulates are likely to dart into the road, you might want to consider a Saab or a Volvo, as their roof pillars and windshields are designed and reinforced with input from Magnus Gens’s moose crash test dummy. Magnus received funding from the Swedish National Road and Transport Research Institute, which shortly thereafter received a plea to help design a camel crash test dummy.

A camel is taller and heavier, and therefore even deadlier, than a moose. More of the roof is likely to collapse directly onto drivers’ heads. If they lean or duck sideways to dodge the hurtling ungulate, now their back is likely to be broken instead of their neck. Of 16 Saudis whose cars struck camels, one study relates, nine wound up with complete quadriplegia. Along certain stretches of highway, camel density has been as high as 19 per mile. These animals are not wild, but their owners often allow them to roam. Sometimes even, in days past, encouraged it. Because until recently, Saudi law required the driver to pay the camel’s owner for the loss. “Therefore,” reports a team of neuroscientists at Riyadh Armed Forces Hospital, “some camel owners have been known to push their animals onto the highways after sunset to claim compensation after the accidents.” A pox on them. Debris, hair, entrails, and faeces on their heads.

Summing up: Do not brake excessively or swerve wildly for a small creature, no matter how cute. Do swerve and brake and run off the road for a camel on an empty desert highway, because there’s nothing to run into but sand. Never speed in moose country. About deer, I don’t know what to tell you. The IIHS study suggests you should brake or swerve only when there’s the space to do so safely, never to the point of skidding or losing control, because deer impacts don’t reliably injure parties other than deer. The alternative is what, ploughing into them? Who does this? People brake. And if they brake hard, the nose of their car drops down and the impact happens lower, perhaps at the level of the deer’s legs, causing more of the torso to pinwheel toward the windshield. And the cars behind to rear-end you. What’s a rational person do?

Let’s ask the most rational driver of all: the autonomous car. If it slams its brakes, does it only do so when no one’s tailgating? If it swerves, does it do so only if the path is clear? If either criterion is missing, will it go ahead and run straight over a beagle or a skunk? I posed these questions to Google/Waymo’s self-driving car media relations person, but she refused to have relations with me. I got no answers and no one to interview, and soon she stopped replying altogether. Somewhere in the middle of our standoff, one of Uber’s autonomous vehicles, travelling at 43 mph, ploughed into an Arizona pedestrian without braking or swerving. As if she were a squirrel. Seems like they don’t have the answers, either.

Posted on Friday, August 18th, 2023 - Categories: #45 - Oct-Dec 2021, App Issues.