What sound level metric to use when relating ambient noise level to animals?

Question

I'm trying to measure elevation in ambient noise levels due to boat activity and relate this to bottlenose dolphins and how it may effect them. Mammals integrate sounds over one-third octave bands (TOBs) and masking occurs when another sound occurs in the same auditory filter band. Therefore, a TOB measurement seems reasonable, although I'm not sure which TOB would be best (e.g. ~100 Hz for boat noise, 2 kHz at the lower range of dolphin whistles). However, boat noise covers may TOBs and bottlenose dolphins generally produce whistles for communication ~2-20 kHz, therefore, a broadband measurement over this range also seems like a good way to relate sound to dolphins.

What would be the best solution? I can see that they may all have different pros and cons.

Answer 1

It depends on whether you want to quantify masking of particular sounds such as whistles or if you want to parameterize loudness to evaluate risk of behavioral responses or, if loud enough, TTS. If you are evaluating masking then consider using a filter that covers the bandwidth of the signals masked: https://marinebioacoustics.files.wordpress.com/2018/04/jensen_et_al_2009b.pdf if you wish to quantify loudness then use audiogram weighting: https://marinebioacoustics.files.wordpress.com/2019/01/tougaard-and-beedholm-2018.pdf

Answer 2

This is a complicated question and one under consideration by much of the field. Spoiler alert, there is no simple solution. Sound can affect animals through 1) masking 2) behavioral disturbance and 3) acoustic trauma.

For boat noise, acoustic trauma is less likely so we won't discuss that right now. That leaves masking and behavioral response. For masking you will want to use the band(s) that cover the frequency range of either the clicks or whistles. This is what's done for killer whales (https://wildwhales.org/wp-content/uploads/2017/09/Read-the-Report.pdf).

Behavioral response is the trickier of the two. Historically, BR models considered only the frequency range over which animals produce sound (typically the most sensitive frequency range) but later studies have shown that animals may react stronger to sounds outside their peak sensitivity. So, for this I would suggest a deep dive into the literature and see what bands (if any) other researchers have used for BR.

As I said and you noted, there is no single answer to your question. I think after a good data dive so long as you can justify the bands you use, you will probably be alright. Good luck!