Developing noise–dosage weighting functions for animals

Human A- and C-weighting functions have been used to calculate noise dosage estimators in studies of wild birds and mammals. It would be better to develop functions specifically designed for them, perhaps taking into account psychophysical differences taxon by taxon. Unfortunately, methods that yield human weighting functions cannot be generalized to most animals. First, human weighting functions are based on equal-loudness contours, but these cannot be measured directly in wild animals. Possible alternate weighting functions include (1) the auditory threshold function of the species/taxon, standardized to 0; (2) the auditory threshold function shifted by a scalar quantity to reflect varying sensitivity; and (3) equal-latency functions, which fit equal-loudness curves well in humans. Second, it is not clear that such weighting functions yield efficient predictors of effect. While human A-weighting is used to calculate noise dosage in studies of both nonauditory and auditory effects, a single weighting function need not yield the best dosage estimator for effects in other animals. In studies of impact on behavior, equal-response-threshold functions (frequency-specific thresholds for behaviors) may be more efficient. Such alternative measures must be compared empirically before dosage estimators can be standardized.