Using echoSOUNDer data from fishing vessels to TRACK ecosystem impacts of the krill fishery

SOUNDTRACK is a project that combines deep learning methods in acoustic data analysis with field sampling and fisheries science toinvestigate the ecosystem impacts of the Antarctic krill fishery and to improve its management. Antarctic krill is a key species in the Antarctic marine ecosystem and the target of a growing commercial fishery. Given krill's ecological importance, concerns have been raised about potential negative ecosystem impacts from fishing. Resource-efficient approaches to support fisheries monitoring are particularly valuable in this context due to the high costs associated with data collection in Antarctica. In SOUNDTRACK, we will first develop a segmentation model to extract dives of air-breathing krill predators, such as whales, seals and penguins, from historical acoustic data recorded by krill fishing vessels during their operations. Acoustic data from fishing vessels are under-utilised, but provide a rich archive of observations of krill swarms and vessel interactions with krill predators. Analysing the frequency of vessel-predator interactions, and their temporal and spatial occurrence in relation to krill distribution and predator demand for krill, will provide new perspectives on the ecosystem impacts of the fishery, using cost-effective data and platforms as a co-benefit. We will complement the analysis of the acoustic data by conducting an ecosystem survey at the Antarctic Peninsula in autumn 2027, coinciding with the fishing season in this region. By combining the scientific survey data with acoustic observations from fishing vessels from the same region, their fleet movement patterns, and the life stage composition of the krill caught, this approach will allow us to provide a detailed analysis of ecosystem impacts of the fishery. The implications of our analyses for fishery management will be communicated directly to CCAMLR, the management body for the krill fishery, and the segmentation model will be provided as a new monitoring tool.