The relative impact of aircraft noise and number in a full-factorial laboratory design

The discussion on the extension of existing airports and the construction of new ones has reached a new peak. This paper reviews the literature on the possibility of trading noise for number of aircraft operations in the context of increasing air travel. Moreover, two new laboratory studies are presented, which model two ways of immission reduction: (1) fading out old and loud aircraft at existing airports and (2) increasing the distance to new airfields. take-off recordings were combined with three numbers of events (3, 9, 27) in nine 27-min exposure conditions. In experiment 1, the different noises were recorded from different aircraft types: A322 with Lmax 81 dB(A), B737 with Lmax 86 dB(A), and MD80 with Lmax 91 dB(A). Experiment 2 used an MD80 take-off recorded in distances of 1000 (Lmax 79 dB(A)) and 500 m (Lmax 85 dB(A)) to the start path, as well as directly beneath (Lmax 90 dB(A)). In both experiments, the 27 min of exposure with 3 loud, 9 medium, or 27 soft take-offs had the same Leq of 70 dB(A) in the exposure room. Each noise–number combination was presented to 12 subjects via loudspeakers. In total, 216 exposure sessions were performed. In both experiments, analyses of variances (ANOVA) revealed the main effects of noise and interactions with number for subjective loudness and annoyance of the experimental noise. The interaction occurred because only three noise events, whether soft, medium or loud, could produce high loudness and annoyance ratings. When subjects imagined the noise being present in the living area, both noise and number had main effects in experiment 1. In experiment 2, subjective responses increased with noise only. Apart from ANOVA, the decibel-equivalent number effect k was calculated. It was expected that k would not exceed 10, which would mean the number of noise events is not considered more than the single noise level in peopleʹs responses to the overall exposure. The analysis revealed k values ranging between −5 and 7.8, confirming that the number variation in the reported experiments affected the responses of the subjects less than the level variation.