Using acoustic monitoring to evaluate the co-benefits of urban restoration

Abstract

Restoring urban greenspaces is an increasingly common nature-based solution aiming to bolster biodiversity while benefiting human wellbeing. However, there is limited understanding of the co-benefits, including potential trade-offs and synergies between outcomes for nature and people. Although passive acoustic recording may be a valuable outcome monitoring tool, appropriate analytical techniques are unknown. We explore how different acoustic analyses capture patterns in species diversity and harmful noise pollution using 2021– 2023 acoustic data from Detroit, Michigan. Data were collected as part of a longitudinal panel study aimed at quantifying the effects of park restoration on physical activity, stress, and cardio-metabolic health. We used AudioMoth recording devices at 100 sites: 10 in the neighborhood around each of 5 restored parks, where native plants were seeded, and 10 around each of 5 unmaintained control parks, matched to restored parks based on specified conditions. Recordings were analyzed by manually identifying bird species and noise in a subset of spectrograms, using an artificial neural network (BirdNET), and using acoustic indices. We found that using acoustic indices in a model could accurately predict bird species diversity and richness, but not community composition. We compared outputs from each type of analysis with mental health indicators of study participants.