Standardised CPUE analysis of the Antarctic toothfish fishery
in CCAMLR Subarea 88.1 from 1988/89 to 2003/04

A toothfish fishery has operated during the Antarctic summer (December - May) from 1997 to 2004, in Subareas 88.1 and 88.2. A preliminary standardised analysis of toothfish CPUE (catch (kg) per baited hook per set) carried out in 2003 compared two alternative toothfish CPUE analyses for the 1998 to 2003 seasons for Subarea 88.1.
This report revises and updates the previous analysis with the addition of data from the 2004 season using data sets based on all vessels and the two main New Zealand vessels that have been in the fishery over the entire time period. However, the fishing grounds were changed to reflect the new SSRUs used to manage the fishery, and the larger regions used to assess the fishery (Dunn et al. 2005). It also examines alternative approaches to modelling catch and effort data by using mixed-effect models following Candy (2003).
The variables entering each of the three models were similar and all explained a reasonable amount (35-46%) of the variation. The variables included in the analysis are plausible, but the models may be influenced by extreme values of soaktime and depth that reflect the exploratory nature of the fishery. Model diagnostics show a reasonable pattern in the residuals for all models, but the quantile-quantile plots indicate a deviation from the normal distribution of the residuals at either end for the lognormal GLM models, and at the lower end for the fixed effects of the mixed model. This suggests that extreme values of catch rate were not modelled well and there may be violations of model assumptions.
The CPUE indices showed consistent trends in all three models. The indices were essentially flat apart from a slight decline in 2001 and a large decline in 2004. The reason for the large decline is unclear, but may reflect the bad ice conditions in 2004 as well as increasing competition between the large number of vessels operating in the fishery.
The relationship between these indices and relative abundance is unknown, and fishery independent data are unavailable to validate this relationship. Continued monitoring of CPUE for the main grounds and main vessels is recommended. Also further research on other possible CPUE models, and inclusion of first order interactions in the analysis is suggested. Research into suitable methods for validating the relationship between CPUE seasonal indices and the relative abundance of toothfish is also recommended.