- The Worldwide Soundscapes project is a global database that documents when and where passive acoustic monitoring has been deployed to study biodiversity in terrestrial as well as aquatic ecosystems.
- The initiative aims to build a community of bioacoustics researchers with the goal of fostering collaboration and facilitating macro-ecological analyses.
- Passive acoustic monitoring, which involves listening in on wildlife using audio recorders, has helped scientists detect elusive species; it also helps understand complicated behavioral patterns.
- Kevin Darras, who initiated the Worldwide Soundscapes project, said the goal is to “harmonize our approaches and to benefit from each other’s experience.”
It started as a simple spreadsheet that documented locations where researchers were recording sound to monitor biodiversity. Three years on, the Worldwide Soundscapes project is a global database on when, how and where passive acoustic monitoring is being deployed around the world to study terrestrial as well as aquatic ecosystems.
“This is a project that is now becoming too big to be handled by only one person,” Kevin Darras, currently senior researcher at France’s National Research Institute for Agriculture, Food and Environment (INRAE), who conceived the project, told Mongabay in a video interview.
Darras started the project when he was a postdoctoral researcher at Westlake University in China. The idea struck when he was waiting for updates on another project he was working on at the time. With the project, Darras said he was attempting to fill a void that often led to duplication of efforts in the research community that uses passive acoustic monitoring — audio recorders left out in the wild — to study biodiversity around the world. “There was a scientific gap in the sense that we didn’t know where and when we were sampling sound for monitoring biodiversity,” he said.
Passive acoustic monitoring has long been used to listen in on insects, birds and other animals in ecosystems around the world. It’s aided scientists to detect elusive species in a noninvasive manner. For example, a team in Australia used acoustic recorders and artificial intelligence to track down the breeding hollows of pink cockatoos (Lophochroa leadbeateri leadbeateri) in a remote region. The method has also helped researchers get insights into the behavioral and communications patterns of animals.
Despite advances in recent years with more sophisticated recorders and automated data analysis, Darras said researchers still haven’t “achieved standardization in terms of deployment or analysis.” Darras said he hoped to use the Worldwide Soundscapes project to help build a supportive network that could potentially work toward harmonizing approaches to passive acoustic monitoring.
“We hope people will look at the data and see what is already done to avoid duplication,” he said. “They might also probably find a colleague’s work and wonder, ‘Oh, why is this gap not filled? Maybe I can do something there.’”
Kevin Darras spoke with Mongabay’s Abhishyant Kidangoor on why he started the Worldwide Soundscapes project, how he envisions it growing into a global network, and the potential of ecoacoustics in biodiversity monitoring. The following interview has been lightly edited for length and clarity.
Mongabay: To start with, how would you describe the Worldwide Soundscapes project to someone who knows nothing about it?
Kevin Darras: In a fairly simple way, I would describe it as a simple inventory of what has been done globally, whether it’s aquatic or terrestrial, in terms of acoustic recording for monitoring biodiversity. Our first goal was to compile something like a phonebook for connecting people who are usually separated by the realms that we study. What I mean by that is we don’t communicate as much among ourselves. For example, marine scientists usually don’t talk much with terrestrial scientists. We have now succeeded in connecting and bringing people together. However, very early on, we realized that we could do more than that, and that we could put our metadata together to get a comprehensive picture of what is going on worldwide in terms of acoustic sampling.
Mongabay: What gaps were you trying to fill with this project?
Kevin Darras: There was a scientific gap in the sense that we didn’t know where and when we were sampling sound for monitoring biodiversity. There was also this gap in the community that made us not so well aware of the developments in other fields. There have been a lot of parallel efforts in different realms when, in reality, the same solutions might already exist in other communities. Our aim is to first make everyone aware of what is out there and ideally, one day, to harmonize our approaches and to benefit from each other’s experience.
Mongabay: Could you give me an example of how acoustic research efforts were duplicated in the past?
Kevin Darras: There are lots of examples when it comes to sound recording, calibration and the deployment of equipment. Because deployment in the deep sea is very much more troublesome and costly, our marine scientists go to great lengths to calibrate their equipment to make every deployment really worth it and to get data that are standardized. As a result, they are able to usually measure noise levels, for instance. Whereas those of us in the terrestrial realm have access to such cheap recorders that setting them up is almost too easy. The consequence is that, generally, we have very large study designs where we deploy hundreds of sensors and recorders and end up with a massive data set that, unfortunately, isn’t very well calibrated. We would only have relative sound levels and won’t be able to really measure noise levels.
On the other hand, I think the community that does terrestrial monitoring has made some great strides with respect to the use of artificial intelligence for identifying sound. By now, we have achieved a pretty consistent approach to bird identification with AI. This is something that could benefit people working in the aquatic realm who often have custom-made analysis procedures.
Mongabay: What was the spark to get started with this?
Kevin Darras: It started three years ago. I was actually busy with another project where I was working on an embedded vision camera. Between the development rounds, we had some time where we were waiting for the next prototype. Rather than just sit and wait, I told my supervisor that I wanted to start another project while waiting for updates. This is when I started contacting people from my close network to find out where they’ve been recording. It started with filling an online spreadsheet, which has grown since then. By now, I believe, a good portion of the community that uses passive acoustic monitoring knows about the project.
Mongabay: Could you tell me how it works currently?
Kevin Darras: The way it currently works is that people find out from their colleagues. Or we actively search for them. Then we send them all the basic information about the project. We ask them to fill in the data in a Google spreadsheet, but we are slowly transitioning to enter everything directly on a website. In the very beginning of the project, we didn’t have the capability, and we needed a really easy and effective way of adding people’s data. A Google spreadsheet was a fairly good idea then. Then we validate the data to see if things make sense. We cross-validate them with our collaborators after showing them the timelines and the maps that represent when and where their recordings have been made. In the end, there is a map which shows where all sounds have been recorded. For each collection, you can also view when exactly the recordings have been made.
Mongabay: Could you give me a sense of the kind of data in the database?
Kevin Darras: If you were a potential contributor, you would have to first provide some general information. Who are the people involved? Are the data externally stored recordings or links? Then we would get to the level of the sampling sites. We require everyone to provide coordinates and also to specify what were the exact ecosystems they were sampling sounds in. That’s the spatial information.
For the temporal information, we ask people to specify when their deployments started and when it stopped, with details on date and time. We also ask for whether they are scheduled recordings with predefined temporal intervals, like daily or weekly, or duty-cycled recordings, meaning one minute or every five minutes, or if they are continuous recordings.
We also request audio parameters like the sampling frequency, high-pass filters, number of channels, the recorders and microphones that they used. Lastly, we ask them to specify whether their deployments were targeting particular [wildlife], which is not always the case. Sometimes people just record soundscapes with a very holistic view.
Mongabay: How do you hope this database will help the community that uses passive acoustic monitoring?
Kevin Darras: We hope people will look at the data and see what is already done to avoid duplication. They might also probably find a colleague’s work and wonder, “Oh, why is this gap not filled? Maybe I can do something there.”
Mongabay: What surprised you the most?
Kevin Darras: It’s probably how big some of these studies were. I was amazed by the sampling effort that, for instance, some Canadian groups did over hundreds of sites over many years.
Also surprising for me was that there were some really gaping holes in our coverage in countries where I would have thought that the means existed for conducting eco-acoustic studies. Many North African countries don’t seem to be doing passive acoustic monitoring. We’ve just had our first collaborator from Turkey. Central Asia is poorly covered. This is for terrestrial monitoring.
For marine monitoring, I was actually surprised to see that the coverage was rather homogeneous. It’s sparse because it’s more difficult to deploy things underwater, but it was globally well distributed. I was surprised to see how many polar deployments there were, for instance, under very challenging conditions. Those are very expensive missions.
Mongabay: What was the biggest challenge in doing this?
Kevin Darras: It’s making everyone happy [laughs].
We had to be fairly flexible with what we expected from people and our criteria. Basically, we decided to trust our collaborators and it worked pretty well. Some people would struggle to provide basic metadata and would have to organize themselves and their data before being able to provide it. Others would be like, “Sure, I can send this to you in five minutes,” and then you get a huge data sheet.
Mongabay: Now that you have a fair idea of how acoustic monitoring is being used around the world, how do you think it is faring when it comes to biodiversity monitoring?
Kevin Darras: I think that the point is too often made that passive acoustic monitoring is something promising and something that has just started. Passive acoustic monitoring has been mature for some time already. It’s true that we haven’t achieved standardization or impact in terms of deployment or analysis, but we are, when using this technology, fairly efficient and effective for gathering rather comprehensive data about biodiversity. I don’t think we need to convince anyone anymore that this is useful and that this is a valid sampling method.
But I have a feeling that this message has not yet reached everyone who’s not using passive acoustic monitoring. It’s rather surprising for me to see that it hasn’t achieved the same level of standardization as what has been done with environmental DNA, when I think that the potential is just as big. Of course, it’s not comparable one to one, but it’s a sampling method that will enable us to have some great global insights.
Mongabay: How do you envision the future of Worldwide Soundscapes?
Kevin Darras: This is a project that is now becoming too big to be handled by only one person. I am soon going to have discussions with the people who want to be involved more deeply so that we have a team that is managing the Worldwide Soundscapes project.
We are going to continue integrating more and more data. We are also looking into automated ways to continue to grow the database from which we can then analyze data to answer macro-ecological questions. As of now, we have only shown the potential of the database. We still need to ask those big ecological questions and show that we can answer them with the database. We would also really like to reach those people in regions where passive acoustic monitoring has not been done yet.
One of the things we’re going to try to develop is something that we’ve tried already on a small scale within our network. To give you an example, I had a North African colleague who wanted to do passive acoustic monitoring in the Sahara and he obtained some recorders from a Polish colleague in the same network. It wasn’t even a loan. They were gifted to him and this enabled him to plug a gap in our coverage. I am hoping that we can develop the network in that sense, where we can loan equipment and provide knowledge for capacity building. It sounds ambitious, but sometimes it’s as simple as sending a postal parcel. I hope it will help expand the use of passive acoustic monitoring.
Abhishyant Kidangoor is a staff writer at Mongabay. Find him on 𝕏 @AbhishyantPK.
Credits
