Résumé
Underwater acoustics have an unrealized potential for multicomponent ecosystem characterization. A variety of methods (e.g., backscattering differences, clustering, Gaussian model) are currently used for multifrequency classification. In this study we propose and implement a new method based on the distribution of scatters on multifrequency spatial planes. Our approach is based on both the sum and the difference of backscattered energy of up to five bi-frequency pairs. This method was developed using four frequencies (38, 70, 120 and 200 kHz) data collected in the frame of the ABRAÇOS (Acoustics along the BRAzilian COaSt) project, around Archipelago of Fernando de Noronha and Atoll das Rocas; an area characterized by low productivity but high biodiversity. By applying the method we could discriminated six groups: Fish like, two types of High Resonant at 38 kHz associated to gelatinous, Fluid Like associated to crustacean macrozooplankton, High Resonant at 70 kHz associated to pelagic algae, and a group of Unclassified echoes. The algorithm was validated using in situ sampling performed by mesopelagic trawl and zooplankton nets. Results are coherent in terms of distribution pattern of each group allows for a 3D representation of organisms distribution around the oceanic islands. Among other, we reveal the importance of gelatinous are the dominant group in the vicinity of the islands where they form a dense layer above the thermocline. These results open new perspectives to improve knowledge on the patterns of distribution and the interaction of a variety of functional groups in tropical and other systems.