The discovery emerged from a passive acoustic monitoring system deployed on a static gill net off the coast of Cornwall. Passive acoustic monitoring, commonly known as PAM, uses underwater instruments to listen to sounds produced by marine animals without actively transmitting signals into the water. Harbour porpoises rely heavily on sound because their underwater environment is often dark, murky and visually limited. They produce extremely high-frequency echolocation clicks, typically far above the range of human hearing, and interpret returning echoes to detect prey, obstacles and other features of their surroundings. The monitoring equipment was designed to study porpoise activity near the fishing gear, but it unintentionally recorded a bycatch event in exceptional detail. The resulting data allowed researchers to compare the animals’ acoustic behaviour with their movements around the net and during the entanglement.

At first, the recordings showed two harbour porpoises foraging close to the fishing gear. Their repeated movements and echolocation activity suggested that the animals were searching for prey in the area while navigating around the net. The porpoises appeared capable of detecting and avoiding the structure for several minutes, a finding that complicates the simple assumption that animals fail to sense fishing nets. Gill nets are made from thin, nearly transparent monofilament or multifilament lines that can be difficult for marine mammals to perceive, particularly when water conditions, light levels or background noise reduce the clarity of the returning echoes. The animals may have recognized the net as a potential hazard, but the researchers believe that a brief loss of attention—possibly during a rapid pursuit of prey—may have brought one porpoise into contact with it. In a moment, an animal that had been successfully navigating around the gear became caught.

Once entangled, the porpoise began a sustained effort to free itself. Movement data indicated that it was able to generate enough force to lift the net, but not enough to tear through or break the material. This distinction is important because the physical properties of fishing gear strongly influence whether an animal survives an encounter. A net that is flexible but exceptionally strong may hold an animal in place while allowing it to drag or raise a section of the gear, increasing energetic demands and restricting access to the surface. Harbour porpoises must breathe air, so every dive is constrained by the need to return to the surface. An entangled animal can become trapped below the water, lose efficient swimming ability, exhaust its oxygen reserves and drown. Even when it can reach the surface, the additional drag and stress may prevent normal breathing, escape or feeding.

The acoustic recordings provided evidence that the second porpoise remained nearby during the crisis. Both animals continued to produce echolocation clicks, which would have helped them sense the net, the surrounding seabed and one another. However, the researchers also detected rapid sequences of clicks that sounded like a buzzing signal after the recordings had been shifted into a frequency range audible to humans. Porpoise clicks are normally ultrasonic and cannot be heard directly by people. The rapid click trains identified in the recording are thought to represent communication signals, potentially associated with distress, agitation or attempts to maintain contact. Although scientists cannot determine the precise meaning of the signals from a single event, the contrast between ordinary echolocation and unusually rapid acoustic activity offers a rare glimpse into how porpoises may respond socially to danger.

The event lasted several minutes, long enough for the instruments to capture changes in movement, sound production and interaction between the two animals. Such data are extraordinarily difficult to obtain because researchers cannot safely or ethically recreate an entanglement, and direct observation in the open ocean is nearly impossible. Monitoring only a fraction of the thousands of kilometres of fishing nets deployed in UK waters, researchers say the probability of recording a bycatch event at the exact moment it occurs is extremely small. The observation therefore acts like an underwater black box, preserving a sequence that would normally vanish without a trace. It shows the animal approaching and working around the net, becoming trapped, attempting to escape and remaining acoustically connected with its companion. For conservation scientists, that sequence is more informative than a simple record that an animal was found dead or missing.

The findings could influence the design of fishing gear and acoustic deterrence technologies. One possibility is to make nets more acoustically reflective, increasing the strength or clarity of the echoes returned to a porpoise and making the net easier to detect. Researchers are also investigating modified net materials and construction methods that could allow larger or stronger animals to break free rather than remain trapped. These approaches must be tested carefully because a change intended to protect porpoises could alter the net’s performance, reduce target catches or create new hazards for other species. Another potential strategy involves biomimicry: developing warning signals inspired by naturally meaningful sounds or by the acoustic cues marine mammals already use to assess danger. The challenge is to discourage porpoises from entering a hazardous area without causing harmful stress, disrupting essential behaviour or driving animals away from important feeding grounds.

The study also highlights why practical solutions must be developed with fishing communities rather than imposed without considering the realities of working at sea. Static net fishing supports coastal livelihoods and is an important part of the fishing industry, meaning that eliminating all such gear is neither a simple nor necessarily realistic conservation strategy. The more immediate goal is to understand precisely how and why entanglements happen, then reduce the risk through evidence-based changes that fishers can use. Researchers at St Andrews work with fishing communities through the UK Bycatch Monitoring Programme, funded by the Department for Environment, Food and Rural Affairs, and through CIBBRiNA, a European Union-funded initiative focused on reducing bycatch of endangered, threatened and protected species. Collaboration can help determine which modifications are affordable, effective and compatible with real fishing conditions.

The scale of the problem makes those improvements urgent. In the United Kingdom alone, approximately 1,000 harbour porpoises are estimated to die each year after becoming trapped in static fishing nets, while global losses affect numerous species of dolphins, porpoises and whales. Harbour porpoises are small, fast-moving cetaceans with high metabolic demands, so an entanglement can become fatal quickly. Their dependence on high-frequency echolocation also means that the acoustic properties of fishing gear may be central to both the problem and the solution. The new recording does not provide a complete explanation for every bycatch event, and the meaning of the porpoises’ communication signals remains uncertain. It does, however, transform an invisible conservation threat into a documented behavioural sequence. By listening to the final minutes of an entangled animal’s struggle and the possible distress calls of its companion, scientists may be closer to designing fishing nets that marine mammals can detect, avoid or escape—turning a tragic accidental recording into a potential blueprint for saving thousands of animals.

Subject of Research: Animals, specifically harbour porpoises

Article Title: First detailed insights into harbour porpoise behaviour during a bycatch event

News Publication Date: 26 August 2026

Web References: https://doi.org/10.1098/rsos.260565

References: Royal Society Open Science, “First detailed insights into harbour porpoise behaviour during a bycatch event,” DOI: 10.1098/rsos.260565

Image Credits: Sea Mammal Research Institute/University of St Andrews

Keywords: harbour porpoise, bycatch, fishing nets, passive acoustic monitoring, marine mammals, echolocation, distress calls, marine conservation, gill nets, cetacean behaviour

In a groundbreaking study published in the prestigious Journal of the Acoustical Society of America (JASA), researchers from the Woods Hole Oceanographic Institution unveiled an innovative approach that marries traditional acoustic monitoring techniques with cutting-edge neural network technologies. By utilizing sound, the study offers a promising new lens through which to observe fish activities and behaviors within coral reefs, providing detailed insights that have previously eluded conservationists.

Traditionally, passive acoustic monitoring has been the go-to method for tracking the activity of coral reefs. Researchers typically deploy underwater acoustic recorders that collect sound data over extended periods, often months at a stretch. While existing signal processing tools have enhanced researchers’ ability to analyze extensive sets of acoustic data, they fall short of efficiently identifying specific sounds. To pinpoint interesting acoustic events, researchers have had to sift through enormous amounts of data manually—a process both labor-intensive and time-consuming.

This tedious task of human-led analysis has long been a significant bottleneck in marine science, drawing criticism for its inefficiency and reliance on human labor. Seth McCammon, one of the authors of the study, candidly described this manual analysis as “awful work,” highlighting the monotonous nature of the task. Beyond its drudgery, the technique lacks scalability, rendering it unsuitable for the urgent need to monitor coral reefs in the face of rapid ecological changes attributed to climate change and other anthropogenic pressures.

The new method introduced by the researchers leverages neural networks, artificial intelligence systems that learn to identify patterns in data. By training these networks to automatically process vast quantities of acoustic data, researchers can facilitate real-time analysis, thereby enhancing the monitoring process significantly. The algorithm they developed has been shown to match the accuracy of human experts—deciphering acoustical trends previously identified only through manual analysis—while operating with an astonishing speed that is over 25 times faster.

This advancement opens the door to exciting possibilities in the realm of ocean conservation. No longer tethered to the limitations of human analysis, researchers are now exploring a range of potential applications beyond static recording devices. For instance, McCammon pointed out the ongoing work with his co-author, Aran Mooney, to integrate neural networks into floating mooring stations that provide real-time updates on fish call counts. These innovations could revolutionize pattern recognition in marine environments, enabling researchers to respond more rapidly to changes in populations and habitats.

One of the most intriguing potentials of this neural network technology is its ability to connect specific acoustic signatures to individual fish species. McCammon highlighted the challenge researchers face in associating unique sounds with particular fish, a “holy grail” of marine acoustic studies. Currently, researchers have yet to conclusively determine which species produce specific calls, creating a gap in understanding fish behavior in a broader ecological context. By detecting fish calls in real time, this pioneering technology aims to facilitate further studies that can link sounds to the fish that produce them, enriching our knowledge of marine life.

As the researchers continue to refine their neural network, they envision a future where real-time monitoring of fish populations becomes commonplace. Such advancements hold the potential for immediate ecological assessments, aiding endangered species identification, and fostering timely responses to ecological disasters. With reefs facing unprecedented threats, from rising ocean temperatures to pollution, the technology could be crucial in painting a clearer picture of reef health and enhancing conservation efforts.

This research demonstrates a paradigm shift in how acoustic data can be harnessed to monitor marine ecosystems, moving away from cumbersome manual processes toward a future reliant on sophisticated artificial intelligence systems. The implications of such a shift extend beyond mere data collection; they herald a future in which monitoring and conservation efforts can occur in a more dynamic, responsive manner.

As a next step in this research journey, the authors aim to integrate their findings into autonomous underwater vehicles, further enhancing the responsiveness and intelligence of ocean monitoring techniques. By deploying neural networks aboard such vehicles, it becomes possible to map out biological activity hotspots in real time, allowing for nuanced insights into the behaviors and distributions of various marine species.

The article titled “Rapid detection of fish calls within diverse coral reef soundscapes using a convolutional neural network” presents a significant advancement in marine ecology, revealing that the intersection of technology and traditional research methods can create new pathways for understanding complex ecosystems. This exciting development reflects the growing recognition of the importance of technology in the face of global ecological crises, underscoring that protecting our ocean environments requires both innovation and dedication.

This research not only catalyzes a new wave of acoustic monitoring but also aligns with the broader goal of preserving biodiversity in our oceans. As the world’s coral reefs face a multitude of threats, the adoption of advanced sound analysis technologies will undoubtedly play a crucial role in shaping effective and timely conservation strategies moving forward. Ultimately, McCammon and his colleagues hope their work will contribute to the larger effort of ensuring the health and longevity of these vital ecosystems for generations to come.

The future of marine research looks promising, driven by innovative technologies that pave the way for an era where scientists can understand and protect marine ecosystems with unprecedented precision and efficiency. As researchers continue to push the boundaries of knowledge and technological capability, we are reminded of the importance of adapting our methods to meet the pressing needs of our planet’s biodiversity.

The study emphasizes the necessity of continuing investment in research and technology to mitigate the alarming declines in biodiversity and the degradation of crucial ecosystems such as coral reefs. By providing conservationists with enhanced tools for monitoring fish populations and identifying species at risk, we are one substantial step closer to safeguarding our oceans and the rich life they harbor.

Subject of Research: The use of neural networks for rapid detection of fish calls in coral reef ecosystems
Article Title: Rapid detection of fish calls within diverse coral reef soundscapes using a convolutional neural network
News Publication Date: March 11, 2025
Web References: DOI Link
References: None available
Image Credits: Austin Greene, Woods Hole Oceanographic Institution
Keywords: Coral reefs, Neural networks, Marine fishes, Sound, Acoustic monitoring, Conservation