CHARACTERIZATION OF THE ENVIRONMENTAL QUALITY OF URBAN GREEN AREAS AND ANALYSIS OF AN URBAN REGENERATION PROJECT BY MEANS OF ECOACOUSTICS

Noise pollution is a major environmental issue, representing the second leading cause of premature death in Europe after air pollution, and affecting both human health and wildlife. Urban noise contributes to physiological and psychological disorders in humans and negatively impacts fauna by altering behaviour, fitness, and ultimately biodiversity. In this context, urban green areas may act as refuges from noise, while regeneration projects aim to improve environmental quality and citizens’ wellbeing by replacing impervious surfaces with vegetation. This research explores the role of ecoacoustics as a tool to assess and monitor changes in urban soundscapes associated with green regeneration interventions. Ecoacoustics integrates passive acoustic monitoring, quantitative indices, and machine learning to characterize sonic environments and biological activity. The study focuses on three sites in Milan (Italy): Piazza della Scienza at the University of Milano-Bicocca, monitored before and after a depaving and greening intervention; the peri-urban Regional Park Parco Nord; and the pocket park Vivarium-Bicocca. A core methodological contribution of this work is the development of an audio equalization protocol to reduce biases introduced by different recording devices. Frequency response linearization was achieved by comparing white-noise recordings from multiple recorders with a class 1 sound level meter in an anechoic chamber. The protocol significantly reduced discrepancies in ecoacoustic indices and was validated on both laboratory and field data; the associated MATLAB script was published open access. The post-operam soundscape of Piazza della Scienza was analysed using ecoacoustic indices, BirdNET-based avian richness estimation, and a custom Random Forest sound-event classifier. Results indicate an increase in acoustic complexity, a redistribution of sound sources, higher avian richness, and more temporally distributed vocal activity. Mixed-effects modelling confirmed a relationship between acoustic complexity and biological and anthropogenic sound components. Comparisons among the three sites show that the regenerated Piazza exhibits a park-like soundscape, particularly at ground-level locations directly affected by greening. Overall, this research demonstrates that low-impact urban interventions can significantly improve urban soundscapes and supports ecoacoustics as an effective monitoring and decision-support tool for urban planning and climate adaptation policies.