Short Wave - Detecting Pests By Eavesdropping On Insects

Episode Date: August 30, 2024

From Indonesia to Wisconsin, farmers all over the world struggle with a huge problem: pests. On top of that, it's tough for farmers to identify where exactly they have the pests and when. Reporter Lina Tran from NPR member station WUWM in Milwaukee joins host Emily Kwong to tell the story of how researchers in the Midwest are inventing new forms of pest detection that involve eavesdropping on the world of insects. Plus, hear what aphid slurping sounds like.If you liked this episode, check out behind-the-scenes photos of Insect Eavesdropper experiments in Lina's digital story!Interested in hearing more insect news? Email us at [e-mail]. See pcm.adswizz.com for information about our collection and use of personal data for sponsorship and to manage your podcast sponsorship preferences.NPR Privacy Policy

You're listening to Shortwave from NPR. Hey, shortwaivers, Emily Kwong here. I've got Lena Tran with me. She's a reporter at WUWM, an NPR network station in Milwaukee, Lena. Hey. Hey, Emily. Thanks for having me. Your latest reporting project is about insects, specifically insect sounds.

I'm so excited for this. I love the hidden worlds of the critters beneath our feet. Where does this begin? Our story starts with a young entomologist at the University of Wisconsin-Madison. Her name is Emily Bick, and her whole thing is coming up with better ways to find agricultural pests, pests that can cause serious economic damage. We lose 20% of our crop to these pests, so we should probably be doing something if we want to conserve more land for non-agricultural use. A few years ago, this huge sugarcane farm in Indonesia, almost 100 square miles, comes to her,

with this challenge. Could she come up with a way to directly and immediately detect the moths chewing up their sugar cane, like the ones that are actively eating up their crop? For a hundred square miles of farm, that is a lot of property to monitor. Exactly. So this sends Emily down a rabbit hole, figuring out a way to do direct pest detection without any harm to the crops themselves. She showed me her method in a cornfield in Wisconsin. And we can walk through over here. I was there early in the season, early enough that the corn was just knee high. I'll walk this way. And it is muggy.

All this corn is sweating under the sun, but these are exactly the conditions that Emily is looking for. When the corn is photosynthesizing and chalk full of sugar, she's here to study the pest that eats that stuff up. A feast for pests. Okay, what kind of pests are picking at this corn? It's called corn root worm. An insect that feeds on corn roots, which makes it hard for the corn to get nutrients and prone to falling over. That sounds like a tough pest. Yeah, corn root worm is one of the worst corn pests in the U.S.

It's called the billion-dollar bug for how much damage it causes. And finding these little guys is really tough. You can't confirm they're there until you uproot the corn. These are hard to see, and they're literally underground. And if we dig up the plant, we lose the plant. All over the world, we're pretty much constantly competing with pests like corn rootworm for our food. And Emily says if we could automate finding them, that could result in using pesticides more strategically, less often at just the right time. That would be ideal. So what was her solution?

She has invented what she calls the insect eavesdropper, a device that tracks pests through sound. So today on the show, we're eavesdropping on the world of insects. How researchers in the Midwest are inventing new forms of pest detection and what aphid slurping sounds like. I'm Emily Kwong and you're listening to Shortwave from NPR. All right. Well, Lena, today we're talking about a novel form of pest detection. That involves listening to insects. I've never heard of this kind of approach.

Before we get into it, how does pest detection happen right now in the U.S.? Yeah, it's a whole industry. Often big-scale farms hire consultants who come a few times a season to monitor long rows of crops. One person might be responsible for up to 20,000 acres. That just sounds like a huge job. There cannot be enough people to cover that much ground. So I imagine there's all these pests flying under the radar. And climate change is also bringing new challenges.

Bugs are thriving in warmer weather, expanding their ranges. One of the scientists I talked to says in our super, connected society, it's going to be really important to food production in the U.S. and worldwide to detect pests quickly. So that's the problem. And now I'm going to tell you the story of how the insect eavesdropper got invented. It's like a farming bedtime story. I'm here for it. Okay. So remember, our entomologist, Emily Bick, had this challenge before her. Find the pest in miles and miles of sugar cane on this farm in Indonesia. Right, because moths were eating. up their sugar cane. Yeah. So Emily goes home, thinks about it. She remembers back in the 80s some

scientists used gramophone needles, which are sensitive to vibrations, to listen to insects that tap on plants to communicate. And from there, she remembers that in the 90s, scientists aimed devices called laser vibrometers at plants. Another decade after that, people strapped accelerometers to plants. You know the little devices in our phones and video game controllers that detect motion? Oh, yeah, from the Wii to the PS5, it's all the rage. You got it. So these were all different ways and expensive ways of tracking the vibrations that come from insect plant interactions. Cool.

All right. And somewhere between vibrometers and accelerometers, Emily falls deep in a research hole. I kind of got myself caught in an area where I'm sure I would have been on some government watch list if I had a search list too much. But I basically realized that the way that the spy agencies listen to people is by attaching these things to walls and the walls would vibrate. And then that vibration would be translated into, you know, what did this person say? She's thinking, is there a way to pick up on the vibrations that an insect makes when it's chomping on some leaves or, you know, chewing on roots or boring down a stalk? Right. She could like, theoretically record those vibrations and translate them to know that. the insect is there.

Yeah, exactly. That's the idea. And she figured out an affordable way to get the job done is with contact mics. It turns out musicians use them a lot. It has this little brass disc that sits on a solid object. And every time there's a vibration, the mics register an electrical signal. Now, when corn root worms feed on the roots, the signal, or essentially the vibrations translate from the root system to the stem, and that's where we're picking it up. So we're using the plant as the musical instrument, as essentially the outside of a guitar.

So Emily and a colleague tested out. We clipped it to desks, and we started talking, and we started tapping. And the fact that the desk couldn't hear our voice vibrations, but it could distinctly and clearly hear those taps enabled us to say, okay, let's take this technology and move it forward. That's brilliant. I mean, she figured out with these contact mics kind of an affordable way to capture vibrations and maybe transfer it to pest control. That's so cool. Yeah, so she turned it into the

insect eavesdropper. You know, it has this mic that you attach to plants and listen for insect feeding sounds, which is what was happening that day I was in the cornfield. They were micing up the corn for an experiment with corn rootworm. Okay, so she developed the idea, is experimenting with the tech on actual plants. How does Emily plan, though, on turning this into an off-the-shelf device that grows in farms around the Midwest and elsewhere it can actually use. Yeah, great question. So for this, Emily needed some coding chops. Enter Dev Marotera, a commsai master student at UW Madison.

He designed the first version of the sensor. And Dev is a computer guy. You know, this whole farming thing was totally new to him. This was like my new time working with plants, insects, and everything. Like, this was the first time I knew how to, like, put a plant in a pot, how much water to give. So Dev's first task is programming this mini-computer to start. stop, and save a recording of insects eating a plant. And they just keep going from there, asking these basic questions.

Like, can they record sound from multiple mics at once? And are they sure the sounds that they have are coming from insects and not say someone just walking by? Okay, so before they can even start thinking about, like, pest ID, they need to refine the technology, it sounds like. Yeah. But eventually they do get there. And things actually move pretty fast because there are these machine learning. algorithms already that kind of do similar things. So dev is able to take what's out there and fine-tune it. Basically, there are a lot more image-classifying algorithms than ones that do audio.

So dev converts the sound files into spectrograms, which are just visual representations of sound. Nice. And it's worked pretty well so far. Their findings aren't published yet, but as of now, their algorithm is anywhere from 80 to 96% accurate, depending on the species. For now, they're focusing on key players in the Midwest, like corn root worm, which kind of plays with its food before it eat. Ooh, I want to hear that sound. Yeah, they make this distinctive tapping sound and also a crunching sound. So what's the tapping for? It's actually to figure out the sugar content and the flow them.

Nice. Yeah, that's cool. Yeah, yeah. They're looking for a sweet treat. And some plants also make anti-insect feeding chemicals, and the tapping can help them detect those, too. And where some insects bite and tap, others suck like apes. Do they have like a straw? They have a straw-like mouthpart.

Wow. And Emily says at first they weren't even sure they could hear this kind of thug, but they were pleasantly surprised. And it kind of sounds like you stuck a straw to the bottom, the dregs of a milkshake. And it sounds like a deep slurping sound. Flom milkshake. Mm-hmm. Yum.

Delicious. So they are improving this technology to identify specific insects because, you know, farmers are only interested in certain ones, the pests. Yeah. And along the way, Dev is interviewing farmers to figure out how to make the eavesdropper a product that will really work for the people who grow our food. Right. I mean, this tech, it's so cool. But my big question, hearing about it is, can it be scaled up for mass agriculture? Yeah. And that's the thing. turns out the farmers are really interested in density, how many of those pests there are. Basically, when the density of insects reaches a certain point, that's when they have to spray

the pesticides. Before that, they cannot spray the pesticides. So they daily have to go out and check in the fields for these insects and count them manually. Whereas the insect eavesdropper will process the clips and say, yep, you have corn root worm or whatever the bug is, and this is how much of it. And Emily is working towards one day using the eavesdropper to monitor. monitor pests across the whole region to provide something kind of like a weather report, but with bugs. Did you get a sense, Lena, for how other people in the field see this work? Yeah, I mean, the scientists I talk to seem really excited. They say sound is a totally new direction for pest detection to go in.

Here's Brian McCornack. He is an entomologist at Kansas State University. I would say it's very novel. And what I really appreciate about Emily's approach is she's always thinking about, you know, the bottom line in terms of, of what's this really going to come to in terms of costs when it be implemented? Emily Bick and her team are still working on the data processing pipeline, you know, how the sound files get cleaned up

before they run them through the algorithm. But she's hoping they can get the insect eavesdropper on the shelf in a couple years. She says sound is just a really cost-effective way to understand the world around us. And I think people are just pumped to hear things that we've never heard before. Sounds that take us inside our food.

and maybe inspire us to listen to everything a little closer. Lena Tran from WUWM in Milwaukee. Thank you so much for this story. Anytime. Thanks for having me. All right, short waivers, if you like this episode, check out the beautiful digital story that Lena wrote. We'll link it in our episode notes. Also, consider following our show on whatever podcast platform you're listening to.

That way you'll never miss another episode. And write to us. We love hearing what science is on our listeners' minds and hearing what you want us to cover in future episodes. Our email is Shortwave at NPR.org. This episode was produced by Jessica Young and edited by our showrunner Rebecca Ramirez. Lena and Jessica checked the facts.

Quasi Lee was the audio engineer. Beth Donovan is our senior director and Colin Campbell is our senior vice president of podcasting strategy. I'm Emily Kwong. Thank you for listening to Shortwave from NPR.