Species identification of fish shoals using coupled split-beam and multibeam echosounders and two scuba-diving observational methods

Species identification remains crucial for interpreting acoustic backscatter delivered by active acoustic meth-odologies. The study took place in a Marine Protected Area where highly restricted areas were present such as no take zones. We used an innovative methodology coupling split-beam and multibeam echosounders to detect and classify monospecific fish shoals (i.e. schools or aggregations). Species identifications were realised by under-water visual censuses made by scientific divers. Two experimental protocols, where the divers gave the identi-fications instantaneously thanks to a communication wireframe, were tested: three roving scuba divers locating the shoals or a towed scuba diver directly behind the vessel. Energy responses, 3-D morphological, shape indexes and spatial descriptive variables of multiple independent samples of 4 observed fish species shoals (Atherina sp., Boops boops, Chromis chromis and Spicara maena) were calculated from the acoustic data. According to their behaviour and feeding strategy, significant differences in the acoustic variables were found between species. The combined use of acoustic data from both echosounders significantly improved the fish species classification. They were well discriminated using a Linear Discriminant Analysis (LDA), including for B. boops, C. chromis and S. maena, which were all observed in aggregations. Finally, we used this LDA model to allocate species to un-known shoals monitored by acoustics methods in the studied site, highlighting the interest of our methodology to predict bentho-pelagic and pelagic fish distributions in shallow waters. We suggest that these acoustic methods to discriminate fish species could provide valuable insights for marine management and decision-making.

Automated summary

Scientists combined split-beam and multibeam echosounders to detect and classify different fish species, and found that acoustic data significantly improved the accuracy of species classification. These acoustic methods have important implications for marine management and decision-making.