Page 58 - Summer 2021

NOISE EXPOSURE CRITERIA
for impulsiveness. A high-stimulus bandwidth is corre- lated to some degree with most, if not all, the impulsive features that have been identified in prior studies, includ- ing kurtosis, crest factor, and rise time. Put simply, it is difficult to generate a stimulus that has a relatively high kurtosis, high crest factor, or rapid signal onset that lacks high-frequency spectral content.
Price (1983) demonstrated that the correspondence between the spectra of a finite transient impulse and hearing sensitivity can effectively predict human thresh- old shifts. Studies of chinchillas using impulsive stimuli demonstrate similar auditory and physiological relation- ships related to noise spectrum (e.g., Ahroon et al., 1996). Although we lack such measurements for marine mam- mals, the conceptual basis of auditory weighting and exposure functions is that auditory effects are likely and occur at the lowest exposure levels at frequencies where the hearing is most sensitive (Figure 4) (Finneran, 2015). Presuming that the absence of high-frequency energy is a reasonable proxy for reduced impulsiveness for multiple metrics, evaluating how spectral content changes from source to receiver could provide insight into which criteria (impulsive or nonimpulsive criteria) is more appropriate.
We are confident that there are broad differences in hear- ing across the frequency spectrum for different marine mammal groups. These may result in differences in whether and how impulsive exposures lose those features that define impulsivity. Although direct measurements to make that determination definitively for any group are again lacking, one consideration for reasoned practical assessments is to identify regions of the frequency spectrum where most taxa overlap in terms of best relative hearing sensitivity and are thus most liable to noise impacts.
It is noted that above about 10 kHz, several things occur both in terms of sound propagation and sound reception. First, broadband stimuli are substantially degraded at this and higher frequencies with increasing range, restricting the overall noise spectrum and thus reducing associated impulsive metrics. Second, this frequency region either overlaps or is fairly similar to best hearing sensitivity rela- tively (not absolutely) speaking across all marine mammal taxa (see Southall et al., 2019a, Figures 1-9). Below this region, high-frequency specialists hear relatively poorly, whereas at higher frequencies, low-frequency hearers are less sensitive. It may be necessary (and eventually
possible) to use different spectral bands as proxies for impulsiveness for different groups. But although there are extensive acknowledged data limitations, as a simple, tractable, interim measure, the presence of noise energy at 10 kHz and higher could be applied. Practically speaking, the range at which noise from an impulsive source lacks discernable energy (relative to ambient noise at the same location) at frequencies ≥ 10 kHz could be used to distin- guishwhentherelevanthearingeffectcriteriatransitions from impulsive to nonimpulsive.
To be clear, measurements of noise exposure at various ranges should continue to be made, ideally including the suite of impulsive metrics measured and considered by Martin et al. (2020). Kurtosis is clearly a useful and insightful metric that should continue to be obtained in real-world measurements. The suggestion here is simply to use a readily quantifiable broadband measurement of spectral energy without distinguishing a specific impulsive metric and threshold that cannot at present be sufficiently prescribed. This approach could also yield directly testable predictions using currently available experimental scenarios. Field recordings of impulsive (near-source) noise obtained at ranges where the prop- agation effects have occurred have yielded insightful results when in laboratory hearing studies (e.g., Sills et al., 2017). A comparable paradigm could be applied for laboratory TTS studies using impulsive noise recorded at ranges where energy above 10 kHz was absent to evaluate TTS onset relative to impulsive and nonimpulsive criteria.
Regardless of whether or how regulators or noise pro- ducers consider the interim approach proposed in this article, it should be recognized that the use of impulsive exposure criteria for receivers at greater ranges (tens of kilometers) is almost certainly an overly precautionary interpretation of existing criteria. This interim measure is presented as a simple, reasonable proxy that could help avoid gross misapplication of auditory criteria until more direct empirical audiometric data are obtained.
Synthesis
This article, which considers auditory and behavioral noise-exposure criteria for different marine mammals and noise sources, marks the conclusion of the second iteration of the noise criteria panel first convened in 2003 by Dr. Roger Gentry. Given rapidly evolving issues sur- rounding marine noise and protected species, Gentry
58 Acoustics Today • Summer 2021