A geospatial model of global ambient sound levelsa)

Area of Science:

• Acoustics
• Geospatial analysis
• Environmental science

Background:

• Ambient sound levels are influenced by diverse natural and anthropogenic sources.
• Understanding global soundscapes is crucial for ecological and human health research.
• Existing models often lack a comprehensive geospatial approach to sound level prediction.

Purpose of the Study:

• To introduce Ambient | Global, a novel geospatial model for predicting worldwide ambient sound levels.
• To quantify the spatial, temporal, and spectral patterns of global soundscapes.
• To differentiate between natural and anthropogenic contributions to ambient sound.

Main Methods:

• Developed an ensemble of four machine learning regression models.
• Utilized a geospatial approach, correlating sound levels with geographic features.
• Trained and validated models using data from acoustic measurement sites.

Main Results:

• Successfully predicted global ambient sound levels based on geospatial features.
• Generated maps illustrating worldwide ambient sound levels (L50), distinguishing natural and anthropogenic components.
• Quantified spatial, temporal, and spectral soundscape patterns under various conditions.

Conclusions:

• The Ambient | Global model provides a powerful tool for analyzing and understanding global soundscapes.
• This research enables further investigation into the impacts of human activities and natural elements on sound environments.
• The model facilitates studies on the effects of ambient sound levels on both ecosystems and human populations worldwide.