Capturing the Sounds of Biodiversity
Friday, July 19, 2013
Researchers at the University of Puerto Rico have developed a system that can automatically record and analyze sound clips from the field. Mitch Aide, lead researcher on the project, describes how the system can improve biodiversity monitoring.
IRA FLATOW, HOST:
Welcome back. I'm Ira Flatow. As I say, a screaming volcano is one kind of natural sound. This is another.
(SOUNDBITE OF WILDLIFE)
FLATOW: That's from an evening at the Lake Selva Biological Station from the Costa Rican tropical rain forest, recorded in 2009 as part of a network of microphones that automatically record sounds, the Automated Remote Biodiversity Monitoring Network or ARBIMON. ARBIMON was developed by a group of researchers from the University of Puerto Rico to eavesdrop on the natural world.
And joining me from San Juan is Mitch Aide. He is one of the lead researchers of ARBIMON and professor in the Department of Biology and Tropical Ecology at the University of Puerto Rico. Welcome to SCIENCE FRIDAY.
DR. MITCH AIDE: Hello, Ira. Thank you.
FLATOW: Those are just fantastic. What - how did you get those sounds?
AIDE: Well, it was - it wasn't really that difficult. We put microphones out in the forest. They were hooked up to a solar panel and an iPod, and the iPod was programmed to record one minute of audio every 10 minutes, which gave us 144 recordings a day.
FLATOW: And what made you think about doing that? What gave you the idea to do that?
AIDE: Well, the funny thing is I'm a plant ecologist. I've worked in the tropics over the last 20 years, and we've gotten much better at monitoring the forest. We have permanent plots placed around many countries and many forests around the world. We also have satellite images that help us have a good idea of rates of deforestation and reforestation.
But we haven't been nearly as good at monitoring the fauna, the animals in the forest. And so we decided with a group of students and a professor from computer science to see if we could use new technologies that are available to improve biodiversity monitoring.
FLATOW: Mm-hmm. And so in that sound we just heard is the whole forest, everything that was picked up together.
AIDE: Yes. You know, we're not trying to record individual species. We're trying to capture the soundscapes, the acoustics of the forest in any one moment. And in that recording, you could hear there was a great tinamou, which is an understory bird species. But in the background, there were also the sounds of the howler monkey and lots of insects.
FLATOW: Lots of insects, you know, frogs...
AIDE: Sure. In that particular recording, I'm not sure. In that site in Costa Rica, there are always lots of insects singing along with the birds, the frogs, the monkeys and probably some biologists as well.
FLATOW: Well, we have more sounds. Wed' love to play them. We have another sound that was recorded at Lake Selva in Costa Rica. Let's listen to that now.
(SOUNDBITE OF WILDLIFE)
FLATOW: That sounds like their calls, different calls going out there.
AIDE: Yeah. The main call is a toucan. People would know that as the bird on Froot Loops.
FLATOW: You're right. Or Disney used to use them also. And so is it possible to separate out what all the different sounds are in that - in all these recordings?
AIDE: Well, all of them, I don't think so.
AIDE: But a large majority of them, we've been very successful at distinguishing individual songs of specific species. You know, that was - one of the major goals of the project was to create the software that allows the user, the biologists to go in and create models that can pull out the individual calls of the species that they're interested in.
FLATOW: So the software is the hero here. It can actually...
FLATOW: ...pick out the species.
AIDE: Well, it's not just the software. You always need the biologist that knows the call of their species. One of the things that people always sort of joke is that we're trying to eliminate the biologists. Now we've automated the collection of the data in the field, and we've automated the species identification. But in between those two steps, you need somebody that really knows their organisms...
AIDE: ...and is able to create the models that can work to help automate the identification.
FLATOW: I'd like to bring on another guest who is using that software, the ARBIMON system also in her own research. Laura May-Collado is a postdoctoral research associate at the University of Vermont in Burlington. She joins us by phone from San Jose, Costa Rica. Welcome to SCIENCE FRIDAY.
LAURA MAY-COLLADO: Hi. Thank you for having me.
FLATOW: Tell us about what your research is - you're involved with.
MAY-COLLADO: Well, I'm on the other side. I'm on the water. I am a marine biologist. And we're trying to implement a similar system that our colleagues in Puerto Rico have done it before. So we specifically are interested in giving or following cetacean species, like humpback whales and bottlenose dolphins, in different parts of Central America. But we're also interested in the South, (unintelligible) information about what species of fish are more common. Some of these fishes of fish are prey for the dolphins. And so it gives us a better way of monitoring these animals, which occupy really, really large areas.
FLATOW: So you put the microphones in the water, and you leave them there, correct?
MAY-COLLADO: Yes. So we have a special underwater unit we leased from colleagues in Hawaii, and these units are called EARs, Ecological Acoustic Recorders. And these units can be deployed underwater and stay there for up to a year. We have one already running in the lease, and we're hoping to deploy another one in Bocas Del Toro in Panama. And hopefully, next year, a few more in the Pacific South, Costa Rica and in Panama.
FLATOW: And give us an idea of what kinds of sounds you've captured.
MAY-COLLADO: We get everything. In Belize, for example, we're getting lots of fish. Fish, contrary to what many people think, they're very noisy. They use sound for communication. We also get lots of shrimps (unintelligible) shrimps and lobster sounds. And, of course, we can get - once in a while, we get a manatee calling, and most of the time, we get bottlenose dolphins. This is in the case of Belize. But you can also in other parts. These units have been deployed all over the world where people are using them to study blue whales, minke whales, humpback whales, et cetera.
FLATOW: Are they - does this software - is the software able to differentiate the sound and tell you which creatures are...
MAY-COLLADO: So while we're using - we want to use ARBIMON, as Mitch was explaining, to you the tools that are ARBIMON provide to us because we can store it, upload these recordings, make them available to the public. But at the same time, we can use tools to identify specific species. So we can use for example frequency or temporal panels used in the ARBIMON software to find, for example, dolphins in thousands and thousands of recording. So that saves us a lot of our time and money as well.
FLATOW: Yeah. And what can you do? What is the aim of that research when you're able to identify?
MAY-COLLADO: Yes. In our case, we are particularly interested in learning how human activities affect a cetacean habitat. So we are hoping that we can associate, for example, boat traffic with changes in acoustic behavior of the dolphins. Also, if we can - if we stop listening to some animals in their home range, it may indicate that these animals are leaving their habitat. So it provides us with lots of tools that we can use to propose strategies of management and conservation in this particular aquatic animal.
FLATOW: Well, thank you very much for taking time to be with us.
MAY-COLLADO: Thank you very much for having me.
FLATOW: You're welcome. Laura May-Collado is a postdoctoral research associate at the University of Vermont in Burlington. She was talking to us from San Jose, Costa Rica. Our number: 1-800-989-8255 if you'd like to talk about things you can hear, things that are going on in the jungles of Costa Rica or Puerto Rico. We're talking about what sounds that are there with Mitch Aide. Mitch, we have another clip I want to play, some more sounds. Maybe you - we'll play it and you can tell us what we're listening to. Let's play that one now.
(SOUNDBITE OF WILDLIFE)
FLATOW: Now, Mitch, what are we listening to there?
AIDE: Well, again that's the toucan in the forest in Costa Rica, the chestnut-mandibled toucan.
FLATOW: I mean, no other animals in there.
AIDE: I'm sure there - again, there's lots of insects that are yet to be identified.
FLATOW: Mm-hmm. Do you ever get any wild boars, peccaries, things like that?
AIDE: Yeah. I think in this recorder - in recording, there might be a bark of a peccarie.
AIDE: People might know it as a javelina. It has a range that extends all the way up into Arizona, New Mexico, and Texas.
There it is. Almost sounds like a bark - a dog barking.
FLATOW: Yeah. Yeah.
AIDE: But that's a...
FLATOW: I thought we had the wrong tape on there for a second.
AIDE: No. No.
FLATOW: Yeah, it was like a dog. It was a dog barking. Do these tell you anything about the health of the animals or the birds out there? Or are they just to collect the sounds?
AIDE: Well, the health of the individuals calling, no. But the idea is - the main idea behind this is to have good long-term data to follow the changes in the populations over time.
AIDE: And we know that climate change and habitat degradation can change the distribution in an abundance of species. The problem is that we just don't have the long-term data to say how these factors are affecting it. And so if we know that a species is frequently calling at a certain time of day and we've seen that pattern for years and then it slowly declines, that gives us an idea of what's happening more at the population level of the species, rather than the health of individuals.
FLATOW: Let's go to the phones now to Davis, California. Let's go to Phil. Hi, Phil. Welcome to SCIENCE FRIDAY.
PHIL: Thank you very much, Ira.
FLATOW: Go ahead.
PHIL: I was wondering if the software has been used to monitor amphibian population. Because I was amazed when I was in Costa Rica how individual the characteristic sounds of the individual frogs were. And, you know, because those amphibians are disappearing.
AIDE: Yes. Mm-hmm. Yes. We've recently published a paper on amphibians here in Puerto Rico. The first permanent station that we established was back in 2008. So now we have over five years' data. And we put it in a wetland in Puerto Rico where a new species of Coqui - this is the common frog in Puerto Rico - but a new species had been discovered.
And of course there wasn't very much information on this species and we established the station there with it collecting 144 recordings a day over a five-year period. And at this site there were at least six different species of frogs that are calling and often they're calling simultaneously. But given that they use different frequencies for their call - they don't overlap - then we were able to create models for each of the different species and come up with good six years' data of how populations have changed across time.
FLATOW: That's great. I'm Ira Flatow. This is SCIENCE FRIDAY from NPR talking with Dr. Mitch Aide, University of Puerto Rico. Let's go to the phones, to Eric in Madison. Hi, Eric.
ERIC: Hi. I had a question about if there's been any progress in identifying to the species level of bats in the neotropics. I know that when I was studying, researching bats in Trinidad and Tobago, some of the anabas detectors and the frequencies that we were getting, we were only really able to identify two genus and not to individuals species. And so I'm curious what the - if the guest knows of any advances...
ERIC: ...in the technologies to get to species levels, you know, the large diversity of bats in the tropics.
FLATOW: Yeah, good question.
AIDE: Yeah, thank you. Our focus has been - has not been on bats. Most bats are calling at higher frequencies than what we're monitoring with the microphones that we're using. In some cases we get bats at lower frequencies and we have been successful in modeling those particular species. But it's something that we would like to do in the future but just getting started, it's kept us very busy just trying to work with birds, amphibians, monkeys, and insects.
FLATOW: Yeah. Let's go to Rob in Rexburg, Idaho. Is that right, Rob?
FLATOW: Hi, go ahead.
ROB: Thanks for taking my call.
FLATOW: You're welcome.
ROB: I am a plant biologist by vocation by my avocation is birding. And so I was wondering with all of the apps that are available for Smartphones these days that oftentimes include the calls, if there was any talk or the potential to incorporate this into sort of a layman's kind of program to help identify birds for us. And again, I appreciate you taking my call and I'll take my answer off the air. Thank you.
FLATOW: Thank you.
AIDE: Right now this project is, more than anything, a proof of concept. We receive money from the National Science Foundation and there we have shown that we can collect data or automate data collection and that the tools that we've developed for automating species identification are working well. And now the project is in the phase of trying to attract more users into the system and expand with the tools and the information that we have.
And one of the things that we've always talked about and wanted to do is get birders involved in this project. We haven't set that up yet but that's one of our first priorities. Because the idea is hopefully to have these permanent recording stations around the globe collecting data, you know, thousands of recordings. It's impossible for us to go through all those recordings but I think it would be something that could be very fun and interesting for birders to listen to.
FLATOW: And crowd source it. Yeah.
FLATOW: You can help them do - and we have a lot of birders who listen to this program. So you're going to get - you will get some response. Before we go I want to - is there a site on the Web that we can all go listen to these sounds on our own?
AIDE: Yes, there is. It's called ARBIMON.net. A-R-B-M-A - A-R-B-I-M-O-N.net.
FLATOW: And if you couldn't write that down, you can go to our website at ScienceFriday.com and we have the link up there for you. It sounds exciting. I'm sure you'll get lots of reaction from...
FLATOW: ...all the birders who listen to us. Thank you very much, Dr. Aide, for talking with us.
AIDE: Yeah. Thank you.
FLATOW: Mitch Aide, the researcher on ARBIMON, Professor in the department of biology and tropical ecology at the University of Puerto Rico. Transcript provided by NPR, Copyright NPR.View this story on npr.org