Anyone fortunate enough to dive on a healthy coral reef will immediately notice the array of noises created by its inhabitants. Pops, crackles, crunches, chirps, hums and snaps produce a vibrant cacophony of sound that indicates the reef is teeming with life. Venture farther out into the ocean and you might hear the clicks, whistles and chattering of a pod of dolphins or the echoes of an intricate whale song. Sound is crucial to life underwater: many species rely primarily on hearing to cue essential behaviors. These include foraging, communication, navigation, hunting, orientation, reproduction and predator avoidance. However, anthropogenic (human-created) noise has become ubiquitous in aquatic ecosystems and is drowning out the natural symphony of sound that so many species depend on.
The underwater growls of huge ships, screeches of military sonar, clattering of pile-driving construction and blasts of seismic survey airguns now pervade the seas. Other noises include the scraping of seafloor dredges, the drone of offshore wind turbines and explosions from dynamite fishing. Sound propagates 4.5 times faster through water than air, and human-generated noise can travel hundreds of miles from its source.
Anthropogenic sound has even penetrated the deepest ocean trenches, making this pollutant both omnipresent and inescapable. It poses a serious threat to aquatic wildlife, yet remains one of the least controlled pollutants on the planet and relatively little is known about the full extent of its impacts.
I first became aware of aquatic noise pollution during lectures in my third-year Oceans module while undertaking my Biology MSci degree at the University of Bristol. The module was a personal favorite — I loved learning about topics such as the origins of oceanic diversity, deep-sea ecology, current marine monitoring techniques and bioluminescence. Another key subject was the human-induced challenges currently faced by marine organisms, including anthropogenic noise.
145 long-finned pilot whales in a mass stranding in Marion Bay in Tasmania, 2005. This was
linked to nearby naval exercises using high-frequency sonar.
Our lectures detailed how anthropogenic noise can impair essential behaviors by masking acoustic signals, distracting attention, eliciting stress responses and even damaging animal hearing thresholds. These impacts can result in reduced foraging efficiency, distorted migration paths, lessened predator avoidance, and diminished reproductive success. In extreme cases, close proximity to anthropogenic noise can impair physiological development, cause lesions on vital organs, and even have fatal consequences. There are distressing examples of this in mass stranding events, where whole pods of dolphins or whales beach themselves. Researchers have linked multiple mass stranding events to nearby naval exercises that involve loud military sonar, which is used to detect submarines. Anthropogenic noise is wreaking havoc on natural aquatic soundscapes, but remains overlooked and understudied.
I quickly became intrigued by this emerging field and was inspired to devise my own research project to test the impact of anthropogenic noise on an aquatic species. There was relatively little research testing how cognition is impacted by noise pollution. This was surprising to me, as cognitive processes are required for actions such as mate choice, foraging, parental care and predator defense, all of which are integral to survival.
Amelia Clarke setting up her detour-reaching experiment in one of the tanks.
One commonly used experiment for testing cognitive performance in non-human animals is the detour-reaching task. It requires the subject to use problem-solving to figure out how to detour around a see-through obstacle to reach a visible goal while inhibiting the natural impulse to attempt to reach it directly. With the guidance of my lecturer and supervisor, Professor Andrew Radford, we conceptualized an experiment testing how anthropogenic noise could impact performance on the detour-reaching task in African daffodil cichlids, small, light tan-colored fish found in Lake Tanganyika.
A group of the study species, African daffodil cichlids, Neolamprologus pulcher.
Daffodil cichlids are highly social and seek the company of other cichlids.
I was awarded generous funding from the Association for the Study of Animal Behaviour to undertake the project during the summer I completed my MSci degree. To the best of my knowledge, no study existed looking at the performance of fish on this particular task when exposed to anthropogenic noise.
The experiment involved placing one focal cichlid behind a liftable opaque partition that was attached to a string pulley system I could operate from behind a curtain. This meant that I could start the trials without the cichlids’ behavior being influenced by my presence. Another cichlid was placed in a transparent holding cylinder behind a transparent C-shaped wall. As daffodil cichlids are highly social and seek the company of others, the cichlid in the transparent holding cylinder acted as a visible social reward for the focal cichlid.
Diagram of Amelia Clarke’s detour-reaching experiment.
Once the experimental trials began, I lifted the partition and timed how long it took for the focal cichlid to solve the task and reach the social reward. I compared the cichlids’ performance under both a silent (control) treatment and a noise treatment played from an underwater speaker. The composition of the noise playback track was designed to simulate the overhead passing of a motorboat, which the cichlids would be likely to encounter in their natural habitat of Lake Tanganyika. To ensure both cichlids felt safe during the experimental trials, I provided each one with a shelter.
However, during these trials, the cichlids displayed little engagement with the cognitive task in either silent or noise treatments. This lack of engagement indicated that my experimental protocol needed refining, so I systematically implemented several changes to my design. The first included increasing the duration of trial times to give the cichlids more time to complete the task. I also decided to settle the focal cichlids in the experimental tank with their bonded partners for a full day to reduce any potential stress arising from separation. I suspected that the cichlids could be displaying neophobia — an aversion to novel objects — toward the C-shaped apparatus. This led to me settling the cichlids with the apparatus over 24 hours.
I also changed the type of shelter provided to the cichlids in the tank from a plant pot to an aquarium plant. I reasoned that the plants would provide a secure base so the cichlids would feel sufficiently safe during trials, but would also leave them exposed enough to entice them to seek the company of the other cichlid behind the C-shaped wall. To gather additional data, I recorded the exploratory behavior of the cichlids in both conditions to determine whether their boldness was impacted by anthropogenic noise. Measuring the cichlids’ exploratory behavior involved measuring how long it took them to fully emerge from the provided shelters when starting the trials.
An African daffodil cichlid about to successfully complete the cognitive task
by detouring around the transparent C-shaped wall to reach a social reward
placed on the other side.
The changes were successful in encouraging greater engagement, which was a satisfying feeling. However, there were still occasions when the cichlids did not engage with the task. I suspect this was due to the social bonds between the focal and reward cichlids not being sufficiently established — meaning the social reward was not a great enough incentive to solve the detour task.
In future studies, allowing more time for these social bonds to be established may provide more motivation for the cichlids to seek the social reward. I carried out my statistical analyses and while some small differences were observed in the exploratory behavior and cognitive performance of the cichlids between silence and noise treatments, these weren’t deemed to be significant. The need to refine my experimental design meant that there was reduced time to carry out testing. It is possible that statistically significant findings would be generated if there were more time to create a larger sample size.
While these results were disappointing, I know that inconclusive findings are an inherent and useful part of scientific research. Rarely are scientific findings perfect and publication-ready on the first try — many are built upon a foundation of previously “failed” experiments. The essential goals of an experiment are to explore a question, to inform future research and to learn something. This project gave me excellent practical experience in devising and modifying experiments, writing successful grant proposals, coding and analyzing large datasets, and handling laboratory animals, so it can hardly be considered a failure. Engaging in this project has further encouraged me to pursue a career in scientific research and, despite the inconclusive findings, I am excited to apply the skills and knowledge I have gained to future projects.
There are many outstanding questions about aquatic anthropogenic noise, including its impact on fish cognition. While short-term impacts have been observed, such as increased stress levels, the long term-effects still remain unclear. Most existing research also concerns individual subjects exposed to one type of noise; further studies are needed to determine the potential population-level consequences and possible cumulative impacts of multiple noise sources. Moreover, anthropogenic noise rarely occurs in isolation to other stressors. Ocean acidification and increased water temperatures impact the transmission of sound in water, so additional experiments are needed to reveal how these stressors could interact with noise pollution in aquatic ecosystems. I look forward to learning more about advances in this field and hope that greater funding and attention are given toward noise pollution research — and that more stringent regulations are placed on the sounds that humans generate underwater.