Transportation noise annoyance: Testing of a probabilistic model

This analysis extends the development and testing of a probabilistic model of noise annoyance proposed in an earlier paper [1]. In the model, noise, measured as the level and number of specific events, leads to annoyance through its interference with activities. Testing of the model is based on noise and survey data collected at 56 sites around Toronto International Airport. Logit analysis is used to estimate equations to predict the probabilities of activity interference and annoyance due to road traffic noise and aircraft noise at each site. From these equations, probabilities of high annoyance are predicted and are compared with the proportion of respondents at each site reporting being highly annoyed. For this comparison, the sites are grouped on the basis of having the same 24 hour Leq. Both predicted and reported annoyance values are consistently higher for aircraft noise than for road traffic noise at the same 24 hour Leq. The results support the underlying argument of the model that previously reported source differences in dose-response relationships can be explained by using a model based on single event rather than daily average noise measures.