Kajsa Rylander-Bellet, 16 May, 2024 | Callan Alexander, School of Biology and Environmental Science, Faculty of Science
Meet Callan, a Threatened Species Project Officer at BirdLife Australia. In 2023, he was recognised for his exceptional volunteer work. But that's just one facet of Callan's dynamic life.
Callan is also a PhD student and Research Assistant at QUT, delving into the intricate world of birdlife in the Granite Belt. His research focuses on how these birds interact with vineyards and the myriad factors influencing their presence, all through the lens of sound. With a background in music and a passion for birdwatching, Callan has found his perfect match in ecoacoustics. In this article, we'll delve into why Callan believes that leveraging AI and technology through ecoacoustics can be a game-changer in species conservation.
Where does your passion for birds stem from?
I was very lucky to have a family that always spent a lot of time in nature. Our holidays were always very nature-oriented. My dad liked taking photos of birds, but I only really started bird-watching in my first few years of university. I found an app called E-bird, which a lot of bird watchers use. It can track all the sounds you hear. This sparked an interest in my brain in how you could track the species that you're seeing and see patterns.
What is your current PhD research about?
My PhD project is looking at bird assemblages on the Granite Belt, with a specific focus on vineyards. The Granite Belt region of southeast Queensland is about three hours west of Brisbane. I focused on Stanthorpe, which has a unique and interesting bird community. Often, we find that rare and unusual birds turn up in the vineyards themselves and almost use the vineyards as habitat. Wine Australia funds my project, and we're looking at how the birds interact with vineyards and what factors influence their presence.
What is bio- and ecoacoustics and what are the advantages of using them?
Bioacoustics and ecoacoustics are often used interchangeably, but there is a difference. Bioacoustics refers to the sound of a living organism for example animal sounds. Whereas ecoacoustics is looking at all sounds in the environment, both manmade, as well as natural.
The advantage of using bio- and eco-acoustics is that you can collect a lot of data in a short time. For example, bird surveying is quite time-intensive, it requires many people out in the field. If you want to survey several sites at once, you need a lot of people, and it’s not feasible. Sound recording enables you to record at the same time every day for months, so you're getting a bird survey every day at hundreds of sites with just one person. The problem is you get an unlistenable amount of data. So, my work is trying to find rapid ways of analysing that acoustic data while still getting meaningful results.
How can machine learning be helpful when using ecoacoustics?
A machine can listen to thousands even millions of hours of data and they are much faster than a human. They're not listening in real-time as a human listen, and they can go through very rapidly because we turn sound into an image, and we train models to identify patterns of images of species that we're interested in.
What's an inspiring moment from your journey at QUT that made you realise how eco acoustics has the potential to save bird species?
One big highlight is the work that we've done on the Easton bristlebird, which is an extremely rare bird. We hadn't seen this extremely rare bird in Queensland since 2010, so we built models to automatically detect the species. There are less than 50 birds in this northern population remaining and we only knew of one population in Queensland. After the 2019 bushfires and droughts, we had no records, so we thought it was extinct in Queensland. But thanks to the acoustic recorders, we've managed to relocate the birds. This is especially useful in the threatened species space. Acoustics can give you a survey effort that you just cannot achieve with ground surveys, because you can monitor crucial sites every single day, which is almost impossible to do with a human, especially in those remote areas.
What advice do you have, who's someone that wants to help save threatened species?
My best recommendation would be to get involved in citizen science. There are lots of citizen science programs that you can do without even having to sign up for a formalised volunteering program. So, things like getting on E-bird, I naturalists record the birds that you're seeing, take photos of things for I naturalist, that's the first step. Then, you can start volunteering for local organisations and use your skills in a way that will benefit conservation outcomes. Everyone has something that they're good at. You might not be a person who can do bird surveys, but you can help in another way.
During your time with Cornell Lab, you wrote many species descriptions for the Merlin Identification app. What is the app, and how can we use it?
The Merlin app is free and built by the Cornell Lab. It can be used to identify photographs of birds you see, and it’s also a free bird guide. So, you don't have to pay to get a bird book or a bird app, it has all the Australian species in it, with photos and sounds. What they're also working on is an automatic sound ID. So, it's like a Shazam for birds. So, if you hear a bird calling, you can push the record button, and it will tell you what bird it is. It's not live yet in Australia but it's coming soon.
For somebody who's experienced both undergraduate and postgraduate studies at QUT, what challenges have you faced and what's your top tip for someone looking to go on a similar journey?
My biggest challenge was just not knowing what I wanted to do, or even knowing that some of these pathways existed when I was in high school, I didn't even know what acoustics were or that it was even something you could do. It wasn’t until I started uni that I realised that it existed. My recommendation for people would be to jump into something that you're interested in and get a better understanding of the entire field. For someone who specifically wants to work in this space, I think you should do a science undergraduate degree and see what you like and push in that direction. I would suggest finding something that you love doing is the best.
Where do you see this kind of AI and ecoacoustic research expanding to?
The research is moving so quickly. The difference between now and even two to three years ago is unbelievable in terms of how good the models are and how people who aren't computer scientists now have much better access to using these models. It's moving so fast that QUT has just partnered with Google on a project where they built a search engine for sound; it's called the Acoustic Observatory Search Tool. So, there's all sorts of amazing things happening in this space, and I think it's just going to keep moving quickly.