Listening lab leading soundscape analyses for United States national parks

UNIVERSITY PARK, Pa. — The Department of Recreation, Park, and Tourism Management in the Penn State College of Health and Human Development recently received a grant from the National Park Service to launch the Protected Areas Research Collaborative (PARC) Listening Lab. It is the sole lab assisting the National Park Service Natural Sounds and Night Skies Division with inventorying and monitoring sounds collected in national parks across the United States.

Increased levels of human-made sound could impact the mating, habitat patterns and feeding habits of wildlife. Run in partnership with the Graduate Program in Acoustics in the Penn State College of Engineering, analyses from this lab help scientists understand how changes in natural or human-made sound affect the quality of visitor experiences and the health of ecosystems. This lab’s work also helps parks identify their baseline level of sound and how it may change over time.

Principal investigators for the PARC Listening Lab include Peter Newman, co-director of the PARC Listening Lab, Martin Professor of Recreation, Park, and Tourism Management and head of the Department of Recreation, Park, and Tourism Management; Derrick Taff, co-director of the PARC Listening Lab and associate professor in the Department of Recreation, Park, and Tourism Management; and Andrew Barnard, director of the Graduate Program in Acoustics and professor in the College of Engineering.

“Population growth and changes in development patterns create human-caused pressure on protected area boundaries, and we are bound to have more interference from human-made noise within park settings,” said Newman, who has worked with the Natural Sounds program of the Natural Sounds and Night Skies Division for 20 years, leading to the creation of this lab. “We need to understand how ecosystems across the globe are changing, and understanding base-level natural sounds helps do that. Once we have a baseline, we can understand how human-caused noise affects ecosystems around the world.”

Scientists at national parks across the country set up microphones and record sounds throughout park systems. The recordings are then transferred to the PARC Listening Lab, where Penn State students listen to, analyze and inventory the sounds heard. Once sounds are analyzed, the findings are sent back to officials at respective parks and used to inform management decisions about how best to protect natural quiet.

Data and analyses stemming from this lab will also help visitor management at natural parks, ensuring visitors abide by park policies that protect and preserve natural or cultural resources. Visitor management includes the infrastructure within parks, such as trails and transportation systems, and the education of park visitors about sustainable practices.

“Natural sounds are important to visitor management and the visitor experience, especially for restoration and recovery,” Taff said. “Health and well-being are top motivators for why people go to outdoor parks and recreation spaces.”

National parks have used soundscape analyses over the past 20 years to understand how items, such as wildfires or the addition of rumble strips on roads, impact the surrounding wildlife, according to Morgan Crump, a dual-degree doctoral student in Recreation, Park, and Tourism Management and Social Data Analytics who oversees the lab’s soundscape analyses.

"It’s critical to understand these soundscapes,” Crump said. “This can be impactful to wildlife health but also visitor experience. For so many people, nature is that escape from reality, and soundscapes are a significant part of that.”

Crump leads a team of undergraduate student researchers, who listen to the sounds and analyze audio with a spectrogram, or a visual representation of frequencies. This spectrogram helps the team visually identify sound, even helping spot the difference between a jet and a prop plane.

The data collected by Crump’s team will ultimately feed an algorithm that uses machine learning and artificial intelligence to inform the inventorying and monitoring of sound clips over time. The algorithm construction is being led by Carter Paprocki, a doctoral student in the Graduate Program in Acoustics.

“This work can lead to better and more informed conservation policies to protect the soundscape of wild areas,” Paprocki said. “The algorithm will help the team deploy analyses to more parks, conduct more studies and reduce turnaround time. This will be an adaptive learning model, so it’s never finished training. The more data the algorithm receives, the better it will get at classifying, understanding and annotating.”

Paprocki said the team is currently in the early stages of constructing the model, working on pre-processing data and finding what data extraction methods, such as using images of spectrograms, can best inform the algorithm. Ultimately, this algorithm will help the team identify the sources of sounds within the parks’ ecosystems.

“With help from this lab, national parks can get a better understanding of what is happening in their areas from an acoustics perspective and what has been happening over the last 20 years as they’ve been making measurements,” Barnard said. “This can be used to make better conservation decisions. The real benefit of this is understanding where we’ve been and then making informed decisions on the future of the parks.”