Exploring spatial relationships of soundscape variables in urban areas

Abstract

This work describes the relationship between soundscape and the underlying landscape in an urban context. In a multifunctional urban area in Würzburg, Germany, soundscape data on pecifically defined spatial scales were observed and evaluated using microcontrollers and Artificial Neural Networks. In order to capture the urban ecosystem function of soundscapes, environmental classifiers are placed at remote sensing designated locations for soundscape classification. Sound source types are classified into Traffic sounds, Human sounds, Biophony sounds, nd Geophony sounds. However, biological and traffic noise appear to coexist at a spatial scale. Significant relationships were also found between noise category classification and several indices of landscape spatial structure, of which vegetation density (NDVI), building density (BD), road density (RD), and patch density (PD) were the most indicative. These results could provide valuable insights for soundscape planning strategies and provide a basis to link soundscapes to land use space.