To estimate optimal relative sample size of scientific observer data collected on Antarctic krill commercial fishing vessels, we estimated the relationship between statistical precision and sample size by using variance component analysis. Observer datasets on Japanese krill fishery from 1995–2008 were analysed by using a hierarchical Bayesian model. The models were composed of multistage cluster units (i.e., year, sub-area, vessel, cruise, and haul) based on a state-space model, separating biological process error in the population dynamics from fishery process as observation error. In both krill length and bycatch fish number, the parameters estimated by MCMC hardly show difference among years, sub-areas, and vessels. The potent interaction effect between year and sub-area suggests large spatio-temporal variability in size structure of krill population, which is presumably derived from large variability of recruitment causing difficulty in predicting krill population dynamics. Variances of observer datasets were calculated by the multistage sampling formula with the variance terms derived from the Bayesian model. For both krill length and bycatch fish number, vessel sample size show marked effects on CV, although haul sample size affect CV for only krill length data up to 10% haul coverage. These results suggest that data collection by scientific observers onboard commercial vessels provide an important information for the management of krill resources and Antarctic ecosystem, while we need further discussion about the optimal relative sample size to ensure the statistical precision required for the specific objective of a study with considering the cost of observer deployment.
Sarah Mackey (CCAMLR Secretariat)