A first-principles model for estimating the prevalence of annoyance with aircraft noise exposure

Abstract

Numerous relationships between noise exposure and transportation noise-induced annoyance have been inferred by curve-fitting methods. The present paper develops a different approach. It derives a systematic relationship by applying an a priori, first-principles model to the findings of forty three studies of the annoyance of aviation noise. The rate of change of annoyance with day-night average sound level (DNL) due to aircraft noise exposure was found to closely resemble the rate of change of loudness with sound level. The agreement of model predictions with the findings of recent curve-fitting exercises (cf. Miedma and Vos, 1998) is noteworthy, considering that other analyses have relied on different analytic methods and disparate data sets. Even though annoyance prevalence rates within individual communities consistently grow in proportion to duration-adjusted loudness, variability in annoyance prevalence rates across communities remains great. The present analyses demonstrate that 1) community-specific differences in annoyance prevalence rates can be plausibly attributed to the joint effect of acoustic and non-DNL related factors and (2) a simple model can account for the aggregate influences of non-DNL related factors on annoyance prevalence rates in different communities in terms of a single parameter expressed in DNL units-a "community tolerance level."

Information-criterion based selection of models for community noise annoyance

“…Fidell et al 4 define L ha ð50Þ as the community tolerance level (CTL), or L CT . Let us further define a community tolerance spread (CTS) as…”

“…10 While it could be argued that this line of reasoning involves a series of empirical deductions and thus amounts to an additional (implicit) parameter, for consistency with Ref. 4 we do not treat the estimated noise dose as an adjustable parameter. In any case, use of an independent annoyance curve for each community survey introduces many new, community-specific parameters.…”

“…This paper examines models for the relationship between DNL and annoyance, and in particular, support for community-specific adjustments such as proposed by Fidell et al 4 There is no intent to ascertain which models are more sensible from a regulatory standpoint; the focus is solely upon which models are better supported by a) A preliminary version of this work was presented in "Assessment of the community tolerance level as a model for the prevalence of noise annoyance," InterNoise 2012, New York, Aug. 2012.…”

“…1,2 Recent research provides evidence that the prevalence of annoyance depends strongly on the attitudes of a particular community toward the noise producing activity; therefore, the predictive value of the acoustical metrics alone is limited. 3,4 Fidell et al 4 thus proposed a model in which each community has its own characteristic noise tolerance, namely, y ¼ e ÀA=m , where y is the fraction of respondents that are highly annoyed by the noise, A is a community-specific constant accounting for non-DNL factors on annoyance, and m ¼ ð10 L dn =10 Þ 0:3 is the estimated noise dose, which relates the rate of growth in annoyance to DNL.…”

“…Most previous dose-response curves have involved two or more explicit, adjustable parameters (e.g., a linear or higher-degree polynomial). 1,8,9 The model of Fidell et al 4 has one explicit parameter, the estimated noise dose. The rate of growth of community annoyance is assumed to be determined by loudness as perceived by an individual listener, which grows as signal energy to the power 0.3, as suggested by Stevens.…”