Tracking Biodiversity: From the Amazon to New Guinea

This is the second in a three-part series examining how communities and researchers are tracking wildlife across Papua New Guinea’s Managalas Plateau.

PART 01: SUMMARY

In the first instalment of this series, communities and researchers began mapping the rich wildlife of Papua New Guinea’s Managalas Plateau using camera traps and acoustic recorders to establish baseline data for biodiversity credits.

Last July, Brazilian biologist and birder Gabriel Leite got the chance to live out one of his dreams: to see — and hear — a bird of paradise in the wild.  

Leite is the biodiversity science lead at wildlife monitoring nonprofit WildMon, which uses technology to detect species, map their distribution and understand how different factors shape their presence in an area. He’s also an avid wildlife sound recorder with an international ranking on Xeno-canto, a citizen science repository for wildlife sounds. 

It was a long journey from Leite’s home in Manaus, a city in the heart of the Amazon, to Managalas Conservation Area (MCA) in PNG’s Oro Province: fifty hours of international travel, a small plane to the town of Popondetta and a bumpy jeep ride to the Plateau itself.  

In this interview for Forests News, Leite discusses his work on the Plateau, how the local rainforest compares to the Amazon and what technology can reveal about wildlife in remote forests.

Why Papua New Guinea, why Managalas? 

Forests News: How did you end up in Papua New Guinea? 

Gabriel Leite: My team at WildMon is responsible for assembling, uploading and analysing data collected from the camera traps and sound recorders. In some projects, especially in a new environment where we [the team] don’t know the forest so well, one of us goes to the field. 

I had never been to Papua New Guinea (PNG) before. It was a long-held dream because of the birds of paradise: I’m an ornithologist, and I did my master’s and PhD on birds. So I always wanted to go, but I’d never had the opportunity or the budget. 

Fieldwork on the Plateau 

FN: What did your days in Managalas look like? 

GL: The main goal was training: working with partners from Landscape Alliance and with local young people — the Biodiversity Team Leaders — who are responsible for deploying the devices. Many of them had little experience using smartphones or the apps we rely on. 

I also spent time in the forest recording species and getting to know the terrain better. That’s really important for our analysis work later on. There are very few recorded examples of things like birdsong from Papua New Guinea, because not many people have gone there to record species.

FN: How easy or difficult was it to access the forest areas you were working in? 

GL: We had to visit each village first and talk with the leaders to secure permission to enter the forest. The places where we were working covered a three-by-three-kilometre area that spanned multiple family properties. But after everything was discussed, the landowners allowed us to go there.  

Finding — and not finding — species 

FN: What surprised you most about the species you encountered? 

GL: I had high expectations, and I was surprised by how difficult it was to see mammals. To find animals like tree kangaroos, you really have to go far from the communities. 

For me, that was a really big contrast. In the Amazon, the forest structure is quite similar, but you see and hear many more species because people rarely hunt, focusing on only one or two specific species. In Managalas, people often hunt anything they can find, even smaller animals. That was a shock for me, and unfortunately, I observed similar patterns in other forest areas I visited in PNG. 

FN: Were the locals able to help you identify many species? 

GL: Mostly just the common ones. There are a lot of parakeets around, which were easy for them to identify. But because hunting is widespread, many other species are scarce and difficult to see. 

One of the guys who went into the forest with me, Reckson Kajiaki, knew quite a lot about the game birds that people hunt. At one point, a bird started calling, and he told me to record it. Then he pointed to it — it was a King bird of paradise [Cicinnurus regius]. That moment was amazing.  

FN: Do you have a sense of how many species you covered, and how many might still be missing from your systems? 

GL: We recorded 130 species on our trip, and collected sound recordings from 70 of them. But of course, there are many more species there than what we observed. The camera traps will help us to identify more of the nocturnal species, and there are likely to be many more species in the highlands, which are harder to get to and where we didn’t spend much time. 

Technology and its limits 

FN: What happens to the data that gets collected? 

GL: We use artificial intelligence(AI) because there’s so much data that it would be impossible for us to go through it all manually. For the camera traps, the program separates the data out into broad groups — sometimes down to family or genus, depending on the photo quality.  

For the sound records, we have another model that detects the calls and matches them to our list of about 400 bird species. But that doesn’t always work, because we don’t have recordings for a lot of species in PNG. If a species is not in the model, we look for other ways to identify them, like asking local people or using other platforms. At this point, we can’t do it all through AI models—my team needs to check everything along the way.

Once we know what species are there, another team does the next level of analysis. They look at different variables like altitude, foliage cover, precipitation, and proximity to communities, and they try to work out what the range of each species might be.  

FN: Can this kind of data be used to estimate population density? 

GL: That’s possible for camera traps—you can get abundance data because you can see the number of individuals. But with the audio recorders, it’s not yet possible to get this kind of information, because we can’t tell how many individuals are singing in each record. That is, we can’t separate the songs out and say, “This one is João, that one is Maria”: it might be the same bird singing all the way through.  

Working with communities 

FN: How did you find working with communities in Managalas? 

GL: It was amazing. Before I went, I heard a lot of things like, “Oh, Papua New Guinea is really dangerous, because of all the fights between the clans.” But I never had any problems. The people were really receptive.  

We had a lot of jokes, because I understood nothing of what they said in the local languages, and English was difficult for almost all of us. But it was amazing to go into the forest with the guys, and they’d tell me about all kinds of things, like the different uses for the trees and plants that we found.

Logistically it’s challenging. The communities are isolated, and it’s hard to get there, especially when the rivers are high, and you have a lot of people and equipment: you often spend a full day trying to get to a community because the roads are bad and you get stuck along the way. But it’s all part of the job. 

Often, when we reached the community, we would all stay the night in a big meeting house. I brought my hammock, like I do when I go into the forest in Brazil, and I was the only one not sleeping on the floor.  

FN: And how did the Biodiversity Team Leaders take to the training? 

GL: They already know their way around the forest pretty well, but we trained them to use the GPS on the cell phones, install the devices correctly and log everything using an app called Survey123.  

It was a little bit tricky; it took a day or two for the guys to get the confidence to go to the field. But when the devices were deployed, and the results came back, it was clear that everybody had done a good job. Of course, the younger ones were much faster to learn everything, but in the end, everybody was able to do the deployment. 

---

For more information on this project, please contact Landscape Alliance’s PNG Programme Lead Will Unsworth: [email protected]