Linda V. Martin
Dept. of Biology, Woods Hole Oceanogr. Inst., Woods Hole, MA 02543
Timothy K. Stanton
Peter H. Wiebe
James F. Lynch
Woods Hole Oceanogr. Inst., Woods Hole, MA 02543
Accurate acoustic characterization of zooplankton species is essential if reliable estimates of zooplankton biomass are to be made from acoustic backscatter measurements of the water column. Work on the forward problem has resulted in the identification of three categories of scatterers, represented by theoretical scattering models [Stanton et al., ICES J. Marine Sci. 51, 505--512 (1994)]: gas-bearing (e.g., siphonophores), fluidlike (e.g., euphausiids), and hard elastic-shelled (e.g., pteropods). If there are consistent differences in the characteristic acoustic signatures of each of these classes of zooplankton, it should be possible to solve the inverse problem by using acoustic backscatter data to mathematically infer the class of scatterer. Two different inversion techniques are applied to a dataset of several hundred pings collected from broadband insonifications (~350--750 kHz) of live zooplankton tethered and suspended in a large tank filled with filtered seawater. The model-based classifiers depend on comparison of the data with theoretical scattering models, whereas the feature-based classifiers are independent of the models and exploit only the inherent characteristics of the acoustic returns. Preliminary results indicate that the acoustic classification of zooplankton insonifications into categories representing distinct scatterer-types is feasible.