Quantifying Area Back Scatter of Marine Organisms in the Arctic Ocean by Machine Learning Based Post-Processing of Volume Back Scatter

As the sea ice reduces in both extent and thickness and the Arctic Ocean opens, there is substantial interest in mapping the marine ecosystem in this remote and until now largely inaccessible ocean. We have used R/V “Kronprins Haakon” during surveys in the central Arctic Ocean in 2022 and 2023, to record the marine ecosystem using modern fisheries acoustics and net sampling. The 2022 survey reached all the way to the North Pole. In a first, rather manually based post-processing of these acoustic recordings using the Large-Scale Post Processing System (LSSS), much effort was used to remove segments of noise due to ice-breaking operations. In a second, more sophisticated post-processing, the KORONA module of LSSS with elements of machine learning was applied for further noise reduction and to allocate the area back-scattering recordings to taxonomic groups as order, families and even species of fish and plankton organisms. We discuss our results with a perspective of underpinning the need for further development of post- processing systems for direct allocation of back-scattered acoustic energy to abundance of categories and even species of marine organisms.