Current krill sampling protocols followed by fishery observers undersample small krill and underestimate the proportion of juvenile krill caught

As an essential part of the krill fishery management imposed by CCAMLR, fishery observers are regularly deployed onboard commercial fishing vessels to monitor the fishing operations. One of their tasks is to take regular samples of the caught krill to record length-frequency distributions and stage composition. These data provide important information on the stage-specific impacts of krill fishing, as well as krill distribution patterns. Scientists from the Alfred Wegener Institute went onboard the Norwegian commercial krill fishing vessel, the Antarctic Endurance, of the company Aker BioMarine during 3 fishing seasons (2020-2022). They conducted extensive length frequency samplings in parallel with the fishery observers. During these samplings, methodological differences between the protocols followed by scientists and observers became apparent. In this study, we compare samplings conducted by observers and scientists to test whether these methodological differences could lead to systematic differences in the demographic characterization of the krill caught. Our results show that the current CCAMLR sampling protocol significantly undersamples the catch fraction of juvenile krill, leading to a biased perception of the stage-specific impacts of the current krill fishery. The main difference between the two sampling protocols is the lack of randomization and the fixed number of individuals sampled in the protocol developed by CCAMLR. Since these data are used by CCAMLR to analyze the age classes of krill that the fishery is harvesting and the impact on local krill populations, we strongly recommend that the current CCAMLR sampling protocol be updated to improve future data quality and ensure adequate monitoring of the krill fishery. These improvements can be easily achieved by incorporating simple steps to ensure random sub-sampling and reduce sampling bias.