Understanding the nonstationary relationships between personal real-time noise exposure and momentary noise annoyance considering various sound characteristics and activity microenvironments

Keywords: noise exposure, mental health, mobile sensing, geographic ecological momentary assessment, uncertain geographic context problem, stationarity bias
Abstract Type: Paper Abstract

Abstract

This study aims to more comprehensively and accurately assess individual exposure to environmental noise pollution by considering various sound characteristics and microenvironments of daily activities, and to further analyze the nonstationary relationships between noise exposure and noise annoyance. We collected real-time data of individuals’ noise exposure and daily activities as well as momentary noise annoyance in Hong Kong using mobile noise sensors and geographic ecological momentary assessment. In addition to the sound level, two novel sound characteristics (sound stability and sound increment within local time windows) were defined and extracted to assess personal noise exposure in actual microenvironments experienced in their daily life. Further, nonparametric classification machine learning models were constructed using random forest to learn the nonstationary relationships between momentary sound characteristics and noise annoyance. The results indicate that (a) the microenvironments of indoor or outdoor daily activities, such as the activity type, travel mode, social interaction and floor, significantly influence personal noise exposure and noise annoyance; (b) different sound characteristics have different effects on individual momentary noise annoyance, and exhibit complex coupling effects with each other; and (c) the relationship between noise exposure and noise annoyance varies over time and the value range of the corresponding sound characteristic.