A geospatial model of ambient sound pressure levels in the contiguous United States

Mennitt D, Sherrill K, Fristrup K. 2014. A geospatial model of ambient sound pressure levels in the contiguous United States. Journal of the Acoustical Society of America. 135(5):2746–2764

This paper presents a model that predicts measured sound pressure levels using geospatial features such as topography, climate, hydrology, and anthropogenic activity. The model utilizesRANDOM FOREST, a tree-based machine learning algorithm, which does not incorporate a priori knowledge of source characteristics or propagation mechanics. The response data encompasses 270 000 h of acoustical measurements from 190 sites located in National Parks across the contiguous United States. The explanatory variables were derived from national geospatial data layers and cross validation procedures were used to evaluate modelperformance and identify variables with predictive power.

Type
Journal Article
Authors
Mennitt, Daniel; Sherrill, Kirk; Fristrup, Kurt
Units
NRSS

Series

Collections

Browse the catalog