Finding a good aircraft noise annoyance curve.

The aim has been to find a good aircraft noise annoyance Dose-Response curve, using practical and robust techniques with the minimum of modelling assumptions.

Several socio-economic/industrial and airport operation factors affected Dose-Response data. This includes ‘population sorting’ at higher noise exposure locations and employment connections, which are likely to reduce annoyance reactions at higher Ldn values; and airport modal effects on people’s recent noise exposure experience, which will produce a defective Dose-Response relationship.

Simple moving average smoothing of the data is a useful procedure. This enables the construction of synthetic large samples – without curve modelling assumptions. It makes apparent the Dose-Response data’s underlying structure. It is straightforward to fit simple curves to this processed data, and to indicate statistical confidence. Note that the large amount of Dose-Response data available is not sampled from a single curve, but rather from a variety of such curves. The assumption is that there is the same underlying ‘mix’ of characteristics in the future.

The analysis has to exclude data from new runways, etc airports. The affected people would be responding to marked Ldn increases over a comparatively short time, not just the actual Ldn at the time of survey. The degree of population sorting is an issue, ie people with high sensitivity moving to a lower noise exposure location.