Welcome to another episode of The Permian Podcast, where we delve into the fascinating world of technology and its applications in tropical forest conservation. Today, we’re diving into the concept of passive acoustic monitoring, breaking down the complexities into simple terms. We’ll also draw comparisons with another part of conservation technology, remote sensing.
Today’s special guest is Albert Aguiar, a biodiversity specialist at Permian Global. Albert has several years of experience in conservation projects aimed at halting bird extinction in the Atlantic Rainforest, Cerrado, Caatinga, and the Amazon.
Albert was also formerly Project Coordinator at SAVE Brasil/Birdlife International with a degree in Biological Sciences (PUCPR), MSc in Animal Biology (UNESP), he specialized in Environmental Management (SENAC) and previously worked as Environmental Consultant.
Mike: Albert, Welcome to the podcast! It’s great to have you on the show today!
Albert: It’s wonderful to be here Mike.
Passive Acoustic Monitoring Explained
Mike: Albert, today’s topic in the field of conservation technology is PAM or passive acoustic monitoring. Can you explain some of the fundamentals of this technology?
Albert: Passive acoustic monitoring, or PAM, or even just acoustic monitoring, is a technique that involves listening to sounds in the environment without actually physically being there. Think of it as kind of eavesdropping on nature! This method is particularly valuable in understanding wildlife, as it allows us to capture the sounds of creatures that communicate using sound, such as whales, or birds, and even bats, that we normally cannot hear on the human spectrum. We also can monitor the soundscape, which means that we could literally see the sound on a broader scale to understand how much life diversity we have in a forest, or in the ocean, simplified to an acoustic index. This index could then be a surrogate for decision making on biodiversity. Of course, all this data can be used for a myriad of scientific exploration.
Mike: Albert, in your case you are working primarily in land-based bioacoustics monitoring, focusing on tropical forest systems I believe…
Albert : Yes! Imagine you have a hidden microphone in the forest, quietly recording the natural soundscape. These recordings can then be analyzed to detect the presence of different species, study their behavior, and even monitor changes in the ecosystem over time. The sound recorders can even act to detect threats to projects, you can even identify chainsaws or shotguns, for instance.
Comparing Approaches – Passive Acoustic Monitoring Vs. Remote Sensing
Mike: What comparisons can we draw with remote sensing, a method that Javier Ramos touched on in a previous episode of the podcast. While both PAM and remote sensing aim to collect data from a distance, their approaches differ significantly.
Albert: Remote sensing often involves the use of satellites or sensors to actively emit signals, like radar or lidar, and then measure the reflected signals to gather information about the Earth’s surface. On the other hand, acoustic sensors don’t emit signals; but rely on ambient sounds for data collection. Basically, an acoustic sensor is formed by a recorder and a microphone. It relies on sound propagation through the air (or water) from the source, an animal for instance, to the sensor. The recorders then transform the sound into electrical signature. Each animal has a unique, or so, signature. We can run analysis based on the frequency and amplitude of the sound, and even see it in a spectrogram, it could be very beautiful, indeed – and allows us to identify what we are hearing but visually. At some point we don´t need to hear, just see to label it is enough when well trained.
Albert: While remote sensing is excellent for mapping large areas and studying surface features, PAM is like having a set of ears close to the ground, allowing us to understand the sonic fingerprints of forest biodiversity in ways that are impossible through observation alone. You need to imagine that each sound recorder is an experienced biologist in the field but not complaining about spending 24-hour periods in the forest for weeks. It is impossible to have 100 biologists in the field, but we can have these sensors that will later be analyzed by AI (Artificial Intelligence) supervised by professionals. This is the reason that we call it passive acoustic monitoring, the sensors are deployed in the field, for months, depending on the specific settings on it. The data could be retrieved from memory cards or even transmitted directly to the cloud. It works in the same way that trapping cameras do.
Remote sensing detailed
Albert: In remote sensing, we primarily rely on monitoring/measuring the electromagnetic spectrum to capture data. Sound is usually not a factor in our studies, as we’re more focused on visual and infrared data. However, there are instances where combining acoustic data with remote sensing can provide a more comprehensive understanding, especially in environmental monitoring. We are now advancing our science by looking at the correlations between soundscapes and biomass from satellite imagery. If you are looking for a part of the Amazon from a satellite, you can say that it´s a pristine and luxurious forest from above. We can then say that the forest is conserved and healthy. However, and I will cite here our experienced colleague here at Permian, Dr. Fabio Olmos, we need to ask the animals if they are happy with the forest. A quiet soundscape might disagree from the aerial footage and its data. The soundscape could demonstrate a history of threats, as we expect more sounds being made by more animals in a conserved patch of forest.
Mike: It’s interesting to see how these seemingly different approaches can find common ground. Javier’s insights highlighted the potential for synergy between remote sensing and passive acoustic monitoring, creating a more holistic approach to environmental studies.
The Future of Passive Acoustic Monitoring
Albert: As technology continues to advance, so does the potential of passive acoustic monitoring. New developments in machine learning and artificial intelligence are making it easier to analyze vast amounts of audio data quickly and accurately. We are not yet in a miraculous scenario where we just plug and get all the answers from the soundscape. Tropical forests are extremely rich and difficult to automate species recognition, take birds as an example, almost 2,000 species here in Brazil, each one, or so, able to emit more than one vocal sign. So, we still rely on experts listening or seeing the recordings and teaching the computers to differentiate the species. While we are not there, there is room to detect a few species of interest, like rare, endemic or threatened ones, that will help us to monitor the project’s impact in the long run.
Albert: Imagine we now have a scenario only predicted to be the future where (low power) microphones are scattered across the forests, continuously listening to the heartbeat of biodiverse life. This wealth of data could revolutionize our understanding of ecosystems and contribute further to our conservation efforts. The data produced could be revisited with deep analysis that still has not been invented yet. We still need to find out what the possibilities are in this field.
Mike: As we wrap up today’s episode, we’ve explored the world of passive acoustic monitoring, breaking down its simplicity and contrasting it with remote sensing. We’ve also seen how these two approaches, seemingly distant in their methods, can complement each other in the pursuit of a deeper understanding of our planet.
Mike: Thank you, Albert, for delving into the details of PAM and taking the time to explain your work in bioacoustics and its contributions to furthering our vital conservation efforts into the future.
Albert: Thank you, Mike!
For more news and podcasts click HERE
Listen on Soundcloud