Comparison of the effectiveness of measures of aircraft noise exposure by using social survey data

Abstract

Results from a social survey of reaction to aircraft noise are used to investigate the effectiveness of various measures of noise exposure. The measures tested include the equal energy index, models involving a logarithmic trade-off of noise level and number of events, “peak-level” models and other non-standard noise descriptors. Analysis methods involve the use of probit analysis to investigate the relative efficiency of the various measures in predicting noise reaction. The results from this investigation indicate that the usual equal energy index gives a prediction of noise reaction which is more accurate than most alternatives, and can be improved only marginally when using the noise exposure variables available in this study. The largest improvement in the predictive power of the equal-energy index is gained by adding to it a linear term in the number of noise events. Time-of-day weightings used in most existing noise exposure measures are shown to be less than optimal.