### Finding Nemo the sprat (and eating him!)

I will continue my slightly creative use of dBSea in this post.

When a porpoise has to echolocate for fish (Nemo) they face a series of challenges, even before taking into account that the fish can swim away.

Harbour porpoises use very high frequency clicks to locate their prey. Kyhn et al. 2013 has some fine data on the clicks they produce, they present the following spectrogram for a Danish harbour porpoise:
 Figure 1. Time series and spectrogram from a harbour porpoise click.
They hypothesise that part of the reason for using such high frequencies (110-170 kHz) is that orcas simply cannot hear them. As porpoises are "snack-sized" for orcas, this might be a good idea.

In dBSea we can model the range a porpoise should be able to pick out for example a sprat. They are a nice size for the porpoises, measuring about 12 cm.
 Figure 2. European sprat, Wikipedia
To find one of these for lunch in murky waters, porpoises listen for the echo from their own emitted clicks. The clicks bounce off of fish in the water, but not all the acoustic energy is reflected. The reflected sound will only be a fraction of the incident sound, the size of this fraction is called the "Target Strength" (TS) and is expressed in dB. For our sprat "FAO" (Food and Agriculture UN) tells us that for a fish of this size the TS is around -49 dB. This means that the reflected sound level of the fish is 49 dB lower than the sound level hitting the fish. In Figure 3(C) the propagation of this echo is plotted (in a quiet water tank). But dBSea can add all sorts of extra depth to our model. In Figure 3(A) the sound field for the sprat in sea state 4. The sound now spreads less far as the model only plots sound levels higher than the noise from the sea state. The last little plot, Figure 3(B), shows the same as A, but taking into account the hearing of harbour porpoises. It indicates that a Porpoise should be able to detect a sprat at 300 metres.
 Figure 3. An "omnidirectional" porpoise transmitting a click like the one in Figure 1. Fish echo A: Echo in sea state 4, B: Echo in sea state 4 weighted for porpoise hearing, C: Echo in a quiet ocean.
I don't think this is the real detection range though, as there are other noises in the sea besides sea state noise. Snapping shrimp makes a lot of noise (great video here or below).

Nevertheless I think this is another interesting way of utilising dBSea for other purposes than strict noise modelling and I sincerely hope that some marine scientist will have a go at something similar, maybe using dBSea?