Can multi-frequency, backscatter data be used to discriminate fish species in multi-species assemblages? It is long recognized that amplitudes of backscattered sound depend, in part, on acoustic carrier frequency. While target strengths from similar sized fish may be the same or different depending on acoustic frequency or species, ratio data or ordinal ranking of echo amplitudes from several frequencies may provide a way to discriminate among targets or even identify species.
As an example, we can examine KRM predicted backscatter from three species of fish commonly found within Lake Michigan: lake whitefish, rainbow smelt, and alewife. Whitefish are much larger and a predator of the two other species. Predicted reduced scattering lengths are plotted as a function of the ratio of total fish length (TL) to acoustic wavelength (l).
The plot shows that reduced scattering lengths differ among species, among fish lengths, and among acoustic frequencies. At 38 kHz amplitudes of the three species are clearly distinguishable. Similarly at 120 kHz echo amplitudes differ but ordinal species rankings differ from those at 38 kHz. Distinguishing species near 200 kHz is more difficult and species rankings change again. No definitive combination of frequencies that universally discriminates species has been identified. The use of backscatter models in comparison with laboratory and field measures is a viable approach to increase the understanding of aquatic organism backscattering characteristics.
|Horne, J.K. 2000. Acoustic approaches to remote species identification: a|
|review. Fisheries Oceanography (in press).|