A storm is blowing in from the west. A helikite shakes in the breeze, and a small group of scientists – dressed mostly in beige with practical zips and stealing worried glances at the sky – start hauling it in, manning the line with heavy gloves like fishermen reeling in the day’s catch.
Butterflies in shades of cobalt and copper flit through the knapweed at the Martin Down Nature Reserve, a 350-hectare stretch of pristine chalk downland in rolling hills just south of Salisbury, in southern England. But there are dark clouds massing on the horizon, and the helikite – a large white helium balloon with a rigid sail to help it take off – strains against its tether, which is attached to a heavy winch anchored to the ground by metal stakes.
Helikites were originally designed for the military. The British Army uses them to carry surveillance gear, and their inventor has pitched them as a low-cost alternative to a US border wall. But they’ve also found widespread use in entomology, where they’re used to take samples of flying insects in large nets tied to the tethering cable at different heights.
Chris Hassall holds the rope as the helikite makes its descent. It’s August 2019, and he is leading a team of researchers from the University of Leeds, using a pair of helikites – the larger the size of a small car – to sample the air at two sites in southern England as part of BioDAR, an ambitious project that could revolutionise our understanding of insects, and help explain why they seem to be dying out.
The basic tools of entomology have remained remarkably unchanged since the Victorian era, when khaki-clad gentleman naturalists descended on distant jungles, sweeping up colourful specimens in butterfly nets, pressing them into heavy books, and sending them to museums in such quantities that many are still sitting in archives waiting to be sorted.
Today’s essential equipment includes pan traps – small bowls half-filled with water and washing-up liquid (to break the surface tension) and painted in ultraviolet hues designed to mimic real flowers (flies like yellow, bees prefer blue). Malaise traps catch creatures flying close to the ground. They look like collapsed tents, with a mesh wall that drives insects into a collecting jar, taking advantage of the tendency of most species to fly straight up when faced with an obstacle. Suction traps sit on long poles and draw in any passing bugs. Sometimes entomologists – usually junior ones – will sit and watch a single plant, tallying all the insects that come near.
And then of course there’s the trusty butterfly net. I join one of the Leeds researchers, Tom Dally – who has short dark hair and a black ear piercing – on a walk around the reserve. He snaps up flying insects in a net with a practiced flick of the wrist: a figure-of-eight motion to first ensnare a creature, and then to flip the net back over the opening to stop it from escaping. He catches things I haven’t even seen. “I can see a half-centimetre long insect from two metres away, but I walk into telegraph poles,” he says.
Each of these sampling methods has its uses. But they have been used sporadically and unsystematically, and don’t paint a representative picture of the whole insect ecosystem. A few long-term studies have been carried out to track insect numbers, mostly by well-meaning amateurs without the scientific rigour required to draw useful conclusions.
“We’re left reading the tea leaves,” says Bill Kunin, an ecology professor at Leeds. Kunin is in his 60s, with wiry brown hair and glasses, and moved to Yorkshire from the US two decades ago after studying at Princeton, Harvard and the University of Washington. He started out in zoology, but drifted towards studying pollinators: the bees, flies and other bugs that are vital to the reproduction of flowers, trees, and the crops that make up more than a third of global agricultural land. In the mid-2000s, he worked on a landmark study that discovered an alarming drop in the diversity of wild bee species in Britain and the Netherlands.
When Kunin was growing up in Minnesota in the 1960s, his parents drove across the Rocky Mountains to Oregon every summer to visit his maternal grandparents. “Every day, the windscreen of the car would just be splattered with dead insects,” he remembers. It was the same in the UK. “If you drove across the countryside, you’d have to wash off your windscreen every day,” he says. “That’s not true anymore.”
Insects seem to be in global decline. Research in Germany suggests the number of flying insects there has fallen more than 75 per cent in the last 30 years. In February 2019, a review of 73 studies conducted since 2013 found that 40 per cent of insect species are in danger of going extinct in the next few decades. The total mass of insects worldwide is dropping by 2.5 per cent per year – a rate that means they could completely vanish within a century.
“It’s important that we understand the decline in insects because they’re absolutely vital to the way that the world works, much more so than all of these lions and tigers that everyone seems to want to conserve,” says Hassall, who is in his 30s, with blue eyes, and a ginger beard. “But the evidence – the quantitative data, the really solid evidence – is almost entirely absent.”
BioDAR aims to change that. Hassall, Kunin and colleagues want to use data from weather radar to identify the signals created by flying insects, and provide some of the hard evidence that’s been missing. “If you can have a bug map that's like a weather map, you can look at where you have hotspots and cold spots and try and relate those back to some underlying causal factor,” Hassall explains.
If it succeeds, the project could lead to the creation of a global map of insect abundance. It could help save endangered species and protect crops in Africa by giving advance warning of swarms of pests. It would be a huge leap forward for entomology. Like all the best ideas, it started in the pub.
In December 2016, Hassall attended a "Crucible" event organised by the University of Leeds to foster collaboration between early-career scientists from different disciplines. Staff members from various departments gathered at Wood Hall Hotel near Wetherby for two days of presentations on topics including public engagement and professional development.
After the talks were finished on the first day, some of the academics congregated in the hotel bar. Hassall and his colleague Liz Duncan, a New Zealander who specialises in the evolution of social insects, got talking to Ryan Neely III, an American radar scientist from the university’s Institute for Climate and Atmospheric Science. Neely, who goes exclusively by his surname (a relic of his days at a boarding school in North Carolina he described as “Harry Potter for science”), has glasses and long brown hair which he ties back in a loose ponytail, and speaks in a quick, southern drawl. He had been in two minds over whether to attend the Crucible event or not. “But I was in a rut,” he says. “I needed to open up and think about things in a different way.”
For meteorologists, insects are noise. Over the years, they’ve developed sophisticated algorithms to sort what they call "hydrometeors" – rain, sleet, snow, hail – from everything else they see on a radar. “We see bees in our weather data,” Neely told Hassall and Duncan when they described their work. “But then we take all that data and we throw it out.” The entomologists looked at each other, slightly horrified. “What are you actually seeing when you say you’re seeing bees?” Hassall remembers asking. “Nobody really knows,” Neely responded. “We just know that it’s not rain.”
Radar, which stands for "radio detection and ranging", relies on bouncing radio waves off distant objects. Differently shaped objects throw varying patterns of this "backscatter" to a receiving dish, and by analysing the reflections it’s possible to work out the shape, speed and direction of things in the sky. The technology was developed in the 1930s, and played a key role in the Second World War, when it was used to track the movements of enemy planes and ships.
But the network of wartime radar stations constructed around southern England also detected weaker signals that radar engineers couldn’t make sense of. These phantom reflections appeared on radar when the skies were completely clear, and moved in seemingly random patterns. The operators dubbed them "angels".
Radar angels may have remained a mystery if not for the efforts of David Lack, a renowned ornithologist who served in the British Army’s Operational Research Group and was stationed in the Orkney Islands during the Second World War, working on applications for radar. Lack thought the angels that his colleagues had been seeing might be birds, and with the help of fellow conscript, entomologist George Varley, proved it – via telescope observations, and an experiment where they strung a dead herring gull from a balloon above a radar station.
After the war, radar became a widespread tool for monitoring weather. But one mystery remained. While it was known that birds accounted for most of the ghost signals, smaller echoes dubbed "dot angels" appeared on some of the new, more powerful meteorological radars being installed across the world. In January 1949, Bell Labs scientist AB Crawford conducted a series of experiments trying to recreate dot angels above a radar station in Arizona. He aimed radar pulses at the exhaust plumes of planes flying overhead, positioned large bonfires so that the smoke would blow into the radar’s path, and set off explosions in the sky. Nothing worked. But when Crawford lined a searchlight up with the direction of the radar at night, he saw small objects moving around in the beam that seemed to match up to the dot angels: insects.
By the 1990s, a new form of radar had been developed specifically for entomology. Instead of a rotating bar scanning an area with the radar station at its centre (a familiar image from old war movies), vertical-looking radar sends a much narrower beam straight up into the sky. “It’s like a searchlight, except that the radar radiation is invisible,” explains Jason Chapman, who has been working with the technology since the late 1990s at the UK’s Natural Resources Institute.
Vertical-looking radar revealed the sheer extent of insect life in the upper atmosphere. “It changed the way people think about insect migration,” Chapman says. “We’ve estimated how many individuals are involved in these migrations and its trillions of insects, thousands of tonnes of biomass, on the move every year.”
However, there are no longer any vertical-looking radars in operation in the UK. “The people who made them have retired, and we don’t have the funds or the expertise to rebuild them,” says Chapman, whose work now involves analysing the data that was collected in the past.
That’s why the BioDAR project hopes to piggyback on more permanent technology – the 18 weather radars that cover the entire UK. “Using the existing weather radar that’s been built for another purpose has pros and cons, but one of the advantages is that there are lots of them,” says Chapman, who has been consulting on the work. “They’re being maintained and we don’t have to worry about them not being there anymore.”
Insects have been seen on weather radar since at least 1962, when a swarm of desert locusts was spotted by a meteorological station near Delhi. In July 2019, the day before I travelled to Leeds to meet Hassall, Neely and Kunin, a swarm of grasshoppers large enough to be seen in weather data descended on Las Vegas. A few weeks earlier, in England, clouds of millions of flying ants were mistaken for rain on the Met Office’s radar systems.
“We’ve known that there have been insects and birds in the weather data,” says Neely. “But we’ve never done anything in a coherent manner.” Until now, the only way biologists have been able to confirm what they were seeing on the radar is through visual observation – essentially, looking out of the window. In the 1950s, ornithologist Sidney Gauthreaux verified nocturnal sightings of migratory birds on weather radar by looking for their silhouettes against the full moon.
In the last few years, however, advances in radar technology have opened up new possibilities. Next-generation dual-polarisation radars send out two sets of radio waves – one set moves in an up-and-down motion, and the other from side to side. Together, they can build a more accurate, three-dimensional picture of the objects that they bounce off in the sky.
For meteorologists, dual polarisation can help tell the difference between sleet, snow and hail. In the hotel bar, Hassall, Neely and Duncan realised they could also use it to identify insects at an unprecedented scale. “Revolutions in radar technology have opened up that new ability to see bugs in a much more comprehensive way,” says Hassall. “We know that you can detect migrating swarms, but that’s not what we’re interested in. We want to get down to the nitty gritty and do this at scale, all the time.”
Several hours – and beers – later, the trio had a plan, involving balloons, blenders and a “biblical apocalypse”.
On a Tuesday in September 2019, Tom Dally boarded a train from Leeds to London with 16 insects slowly defrosting in his luggage. The insects had mostly been collected in spare moments during the fieldwork at Martin Down. Some, such as desert locusts, were bought from local pet shops (where they’re sold as reptile food) – and Dally caught one in the shower of the holiday cottage the group were staying at, charging into the room with a butterfly net at the ready.
When entomologists need to keep an insect specimen for further research, they transfer it to a "killing jar" – a plastic container with a layer of plaster of Paris at the bottom. You put an insect inside, add a few drops of a killing agent such as ethyl acetate, and close the lid. The agent evaporates and kills the specimen as quickly and humanely as possible.
Dally – who seemed genuinely mournful about having to kill anything – took the tube from King's Cross to South Kensington, and spent the evening in a room at Imperial College, extracting the specimens from tissue-stuffed plastic jars and pinning the dead insects to pieces of card. “I dread to think what the housekeepers thought,” he says.
The next day, at the Natural History Museum, the bugs were placed two or three at a time into the vacuum chamber of a machine called a "sputter-coater", which sprayed them with an extremely thin – 20 nanometres – layer of gold palladium. Then each specimen was placed on the rotating turntable of the museum’s micro-CT scanner, which takes incredibly detailed images from every angle. The spray helps to increase the contrast, particularly for fine details such as wings, which are only a few cells thick on some insects. Each scan took 53 minutes.
The team will use the scans – more than 80 in total – to simulate what each of the insects should look like to a weather radar. “We know that there are bugs in the radar data,” says Hassall. “What we don’t know is what each bug looks like in the radar beam.”
First, the CT scans will be used to create 3D models of each insect, in different shapes and flying positions – in what Neely says is more of an art than a science. Those models will then be fed into a software package called WIPL-D, which will estimate what an individual creature ought to look like to a dual-polarising weather radar. It’s normally used to simulate radar profiles for new types of aircraft. “It processes overnight and gives you a beautiful picture of the radiation power bouncing off it,” says Neely. “I think it’s even prettier than the insect.”
Unlike vertical-looking radar, which can draw finer detail by using a much narrower beam, weather radar doesn’t generally have resolution high enough to actually see individual insects. Instead, it sees in "voxels" – the 3D equivalent of pixels – of 100m a side, with all the reflections encountered in that space added together.
To bridge the gap, the projections created by WIPL-D will be fed into SimRADAR, a software package originally developed by researchers in Oklahoma to track the movement of debris inside a tornado. This will require a huge amount of processing power. When I went to visit the team in July 2019, Neely had just ordered a new computer to help deal with all the computation required, and he was very excited about it.
The aim is to create a “biometeor classification algorithm” – they want to be able to look at a patch of clear sky with a weather radar and interpret what would previously have been discarded as random noise. It should be able to tell scientists how much insect mass is in the sky in a given area, and also a sense of what shape the things up there are. “We ought to be able to tell whether these things are long and skinny or short and fat,” Kunin says. “That won’t give us a species ID, but it might give us family-level stuff – are they dragonflies or aphids?”
That information can give a sense of the diversity of an ecosystem, which is as important a measure of the health of the natural world as overall abundance; there may be lots of insects in a given area, but if they’re all the same species, that’s probably not a good sign. “It’s not necessarily understanding which species are there, but understanding how many species there are,” Hassall says.
BioDAR won’t be able to tell whether there’s a seven-spot ladybird at large, but it should be able to create an estimate of the mass of insects in a given patch of sky – every five minutes, anywhere in the UK. But to make sure such interpretations are actually accurate, Hassall and his team have to validate them, which is where the fun really starts.
By the summer of 2017, Hassall, Neely and Duncan had settled on a name for the project (it started off as BeeDAR, but bees tend to fly a bit too low for radar to spot them in large numbers) and secured funding from the university to do a small pilot study.
That work helped to secure a three-year, £600K grant from the Natural Environment Research Council. In summer 2020, aerial sampling with helikites will be used to compare what the radar sees in a block of sky with the insects actually flying in it at the time. “We'll be able to see over the course of a few days what the radar sees in that space, and then we'll be able to sample over a day,” explains Hassall. “Hopefully what we'll see is that we collect more bugs where the radar sees more bugs.”
Each morning, the team will launch the helikite to a height of a thousand metres (they had to get special permission from the Civil Aviation Authority). The biggest version of the kite takes three canisters of helium and about half an hour to inflate – the team tie it down under a tarp overnight so they don’t have to refill it from scratch every day. (Helium is expensive, and there’s a global shortage.)
As well as nets every 200 metres, and flags to warn local pilots, the tether has a series of lightweight sensors carried in polystyrene coffee cups, to measure pressure, temperature and humidity, and make sure the samples are being taken at the right heights. Freya Addison, a physicist and PhD student on Neely’s team, will monitor these from her laptop in real time. If disaster strikes and the tether snaps, there’s also an "automatic cut-down system" built into the helikite – a superheated filament that bursts the whole balloon if it drifts too far away from its launching point.
During the preliminary fieldwork in August 2019, the nets went up at dawn and were pulled in just before dusk each day, weather permitting. Normally, extracting bugs from a net involves a device called a pooter – a plastic jar with two tubes running out of it. You throw the net over your head, point one tube at the insect, and suck through the other to draw it into the jar. A mesh over the sucking tube stops the bug going straight down your throat.
Repeating that procedure for every bug in each of the nets would have taken hours, so Dally improvised a quicker solution – essentially a handheld vacuum cleaner that has been modified to hoover up insects. “Entomologists have a long history of hacking pieces of kit,” says Hassall, as he flicks through images of scientists wearing Ghostbusters-style vacuum backpacks on Google Images. “We’re not that high tech.”
By the end of their week at Martin Down, the freezer of the group’s rented cottage was stuffed with insects. Once they got back to Leeds, the samples needed to be sorted and classified. Some species differ so slightly that even experts can only tell the difference under a microscope. “There are a few that are distinctive enough that you can ID them on the wing, but there are some where the only way you can tell them apart is by dissecting their genitalia,” says Kunin.
That’s one of the reasons why the historical data on insect diversity is so spotty. In 2017, a landmark study by Caspar Hallmann and colleagues reported on the efforts of a group of German amateur naturalists. They had used malaise traps to collect samples in the same areas repeatedly over the course of 27 years. The group gathered so many insects that the only way they could make sense of the data was to literally compare the weight – the biomass – of what they’d collected in each year.
New science offers a quicker way of identifying individual species: metabarcoding. Liz Duncan will be putting some of the insect samples through a machine called a homogeniser – essentially a high-tech blender – which will turn them into what Neely refers to as a “bug milkshake”. This genetic smoothie can then be analysed for DNA markers unique to different species, and therefore determine exactly which creatures went into the original sample. “Flies and moths look the same, and it takes hours to identify them visually,” Duncan says. “But DNA can give you rough abundance and diversity.”
For the final part of the project – which everyone kept referring to as the “biblical apocalypse” – the BioDAR team will be using the equipment at Chilbolton Observatory near Stockbridge in Hampshire, home of the world’s most powerful fully steerable meteorological radar. It can take a snapshot of the sky every six seconds, and is normally used for tracking microsatellites. “It’s pretty intimidating when this thing turns towards you,” Hassall says. “It’s like the Death Star.”
The plan is to breed hundreds of thousands of blowflies – a couple of shoeboxes’ worth – and send them up in a balloon above the radar, then drop them from the sky over Hampshire. “We’ll have to find a very understanding landowner,” Hassall says. “We keep mentioning it to people and they get a little bit tetchy.”
Comparing simulations of what they’d expect to see on the radar as this ball of insects disperses with the actual pattern created by such an event will be the final test for BioDAR. If it succeeds, Hassall hopes the algorithms the team are creating could be exported to other countries, to take advantage of the noise in their weather data that’s currently being discarded. “We could potentially have an almost global measure of insect aerial biomass within five to ten years, and that’s exciting,” he says. “That’s really cool.” And it’s just the start – in Africa, the technology behind BioDAR could have a huge impact on people’s lives.
A few years ago – no one knows exactly when or how – an invasive Central American moth called fall armyworm crossed the Atlantic and appeared in Africa for the first time. It takes its name from its relentless appetite. The larvae advance through fields of maize and sugar cane, leaving plants torn and tattered as if caught in a hailstorm. The adult moths can fly up to 60 miles a night.
“The speed at which it has spread is phenomenal,” says Kunin. Fall armyworm was confirmed in Nigeria in January 2016. In 2017, Kenya lost two months’ worth of maize. By December 2017, it was confirmed or suspected in all of sub-Saharan Africa – a huge problem on a continent where crop pests are already a big issue.
In central Africa, pest infestation accounts for up to 50 per cent of pre-harvest crop loss. In the east, silverleaf whiteflies ravage cassava plantations. Further north, desert locusts explode from the Sahara and strip fields bare. Since 2017, fall armyworm has cost African farmers more than £10 billion, according to a report from the Centre for Agriculture and Bioscience International.
Kunin is leading a sister project to BioDAR, with an 18-month, $100k grant from the Bill & Melinda Gates Foundation. The hope is that the same kinds of algorithms being developed in the UK could be adapted to detect swarms of flying crop pests in Africa.
There are only a handful of dual-polarisation radars in Africa, however. Eight are in South Africa, but most are being run in single-polarisation mode because it’s cheaper. Mali has three, but the British Foreign Office currently advises against all non-essential travel to the war-torn country. So in September 2019, Kunin and Freya Addison flew to Rwanda, in east Africa, which has one dual-polarisation weather radar near Kinazi, about 30km south of the capital, Kigali.
In the Rwandan capital, Kunin and Addison held preliminary meetings for what they’re calling PestDAR, including talks with entomologists from the University of Rwanda and the country’s agricultural board, who are already conducting ground-based sampling work on fall armyworm with pheromone traps across the country. The plan is to combine that information with scans of adult fall armyworm moths and with historical radar data from Meteo Rwanda, the country’s weather service.
There’s a lot of planning work to do, but eventually Kunin wants to create a system that gives farmers advance notice of fall armyworms on the march – so that they can take action by, for example, spraying their crops with pesticides at the most effective time. The BioDAR approach could be applied to all manner of similar challenges, all over the world. “I’m really excited about it,” says Kunin. “It’s something that is much more powerful than anything we’re doing now, and it’s free. The possibilities are only limited by your imagination.”
This article was originally published by WIRED UK
