ASA PRESSROOM


Acoustical Society of America
157th Meeting Lay Language Papers


[ Lay Language Paper Index | Press Room ]


Passive Acoustic Detection of Herring Size

Thomas Hahn - thahn@rsmas.miami.edu
Jennifer Wylie - jwylie@rsmas.miami.edu
Xiaojun Chen - xchen@rsmas.miami.edu
Rosenstiel School of Marine and Atmospheric Sciences, University of Miami

Popular version of paper 1aABa13
Presented Monday morning, May 18, 2009
157th ASA Meeting, Portland, OR

Sound is well known to be the predominate remote sensing tool used to locate and measure fish in the ocean. In the traditional approach, which is similar to using a flashlight in the dark, an artificial sound signal is projected towards potential targets. From the reflected signal one obtains the location of the target, as well as species and size information if the reflectivity of the fish or fish schools is well understood. This principle is implemented in all recreational, commercial, and scientific acoustic fish finders.

In contrast, we attempt to achieve much of these results without using an artificial sound source, but solely using the naturally occurring sound in the ocean. This has the potential advantage that only a very simple piece of equipment, a single micro- (hydro-) phone, needs to be deployed to cover a large area, thus reducing the need for expensive ship operations. In principle, there are many ambient sound sources that could be used for this purpose, including the sound of breaking waves.

In the case of herring, we can use the sound the fish create themselves. Large herring aggregations that form during their spawning seasons can be significant sources of underwater sound. The sound is created when the herring release bubbling air from their swim bladders in response to predatory activity or when adapting to hydrostatic pressure changes that occur during migration from deeper water to the ocean surface for feeding. In a previous work by one of the others, we demonstrated that this gas release produces a typical sound that uniquely signals the presence of herring. Included is a sound clip of the herring response to predatory activity. Figure 1 shows the graphical representation of this sound in the form of a spectrum.

Figure 1. Graphical Representation of the Herring Spectrum

LISTEN: Herring Sound

When a school of fish is exposed to the ambient sound, their swim bladders begin to oscillate and start radiating their own characteristic sound, called the resonance frequency. This alters the ambient sound field in characteristic ways, particularly at the size-dependent pitch of the radiating swim bladders. When a school of fish is located between a source and a receiver, parts of the source spectrum are removed by the swim bladders because sound of this frequency is redirected in other directions. This leads to absorption lines in the spectrum. Likewise, when a fish school is located besides the direct source receiver line, the opposite happens and spectral emission lines appear. An example of a spectral emission line can be viewed in figure 2.

Figure 2. Herring Spectrum with an emission line

Over sufficiently long periods of observation, both spectral emission situations will occur. The net effect is a stronger variability of the spectrum, which can be viewed using statistical analysis, at the frequencies corresponding to fish bladder resonances. If one knows roughly the depths at which the fish were located during the recording time, the size of the swim bladders can easily be inferred using this and the resonance frequency information. For many fish, the size of the swim bladder is a good indicator of overall fish length, which then can be obtained by our new method. In figures 3 and 4, the spectral variability is shown along with a plot of fish length vs. resonance frequency. Using this data we estimated the fish size to be 21.3cm, which was in good coordination with the specimens caught during the recording time.

Figure 3. The Fluctuations of the Herring Spectrum as a function frequency

Figure 4. The Fluctuations (zoomed in on) and an estimation of the fish length for those frequencies. Note the line at the peak fluctuation and the estimation of the fish length of 21.3cm from that peak

[ Lay Language Paper Index | Press Room ]