The effect on annoyance estimation of noise modeling procedures

Worldwide the impact of noise exposure on a community is estimated by the per-centage highly annoyed. The proper determination of the percentage highly annoyed depends among other factors related to the quality of the survey (sample size, selec-tion, questionnaire) also on the quality of the noise assignments related to the type and quality of the modeling procedure. Meta-analyses have demonstrated large variations in the exposure-annoyance curves and described personal, environmental and social factors which may be responsible for these differences (Job 1988; Fields 1993; Miedema & Vos 1999, 2003; Miedema & Fields 2005). Even when these indi-vidual differences are accounted for, Fields et al. (2000) have observed that on aver-age the community response still differs the equivalent of about 7 decibel in noise exposure. At the last ICBEN conference Fields (2003) stated “There is almost no re-search” into these differences. It would, however, be important to investigate the de-terminants of these differences, “because it identifies communities that might be treated differently in noise regulations”. In practice, several modeling standards and minor or larger variations of these coexist. Hitherto, a systematic cross-validation of the effect of the various modeling procedures on the estimation of the percentage highly annoyed in community studies is missing. Realizing this fact, the European community has spent a lot of research money (Harmonoise & Imagine projects) to establish a new modeling standard. This new standard should account better for to-pographical and meteorological conditions which vary a lot across (not