CCAMLR

In 2013, the Scientific Committee of CCAMLR could not achieve consensus on a stock status for Antarctic toothfish in Subarea 88.2 SSRUs 88.2C-H and identified several areas for further work. This was presented to WG-SAM in 2014, and found that the models were unable to fit patterns in the recaptures of tagged fish seen in the SSRU 88.2H fishery. The patterns showed a sharp decay rate of a cohort of tagged fish with few being recaptured after more than 3 years at liberty, a steepening of the decay rate over time, and a trend for increasing proportions of tagged fish caught over time.

To help understand the population dynamics which could explain the observed patterns in the tag recapture data we carried out a series of simple simulations in R. This suggested that the observed pattern for that tagging data in SSRU 88.2H was only reproduced in scenarios that included both immigration and emigration,  combined with moderate to high exploitation rates

Within a single area model, emigration can be mimicked by treating it either as a constant biomass of removals or as an additional mortality rate. Both of these approaches were unsuccessful in achieving an adequate fit to the tag data. Clearly, the steep decline in the recapture rates of a cohort of tagged fish through time cannot be explained solely by emigration (a process that includes both tagged and untagged fish), but requires a significant amount of immigration each year to explain the remaining data. Models that include more than one area may be required to model both immigration into SSRU 88.2H and the subsequent emigration back to SSRUs 88.2C–G.

As requested by WG-SAM we have developed models for (i) SSRU 88.2H using tag recaptures from 1 year at liberty; (ii) SSRU 88.2H using tag recaptures from 3 years at liberty; (iii) SSRUs 88.2C-H using tag recaptures from 1 year at liberty, but excluding tags from the south; and (iv) SSRUs 88.2C-H using tag recaptures from 3 years at liberty, but excluding tags from the south. For each model we provide estimates of biomass and yields based on the CCAMLR decision rules. However, we caution that none of these models provide an adequate explanation of the observations available — none of the model captured either the sharp decay in the number of tagged fish recaptured with increasing time at liberty, nor the steepening in that decay over time.

At its 2011 meeting, the Scientific Committee agreed that a time series of relative abundance from a well-designed survey could be a useful input into the Ross Sea stock assessment model. The first survey was completed in February 2012, and the second survey in February 2013. In this paper we provide results of the third survey in the time series. The objectives of this third survey were: (1) To carry out a longline survey to monitor subadult toothfish in the southern Ross Sea (strata A–C) using standardised gear in a standardised manner; and (2) To sample additional experimental stations in an adjacent area to identify areas of high subadult abundance which could be included as strata in future annual surveys.

The 2014 survey was successful in completing all the planned stations. Standardised catch rates for the core strata showed a slight decline across the three surveys but this decline was not significantly different. Age frequency data from the surveys have shown the progression of a cohort from age 7 in 2012 to age 9 in 2014. These results suggest that the surveys are indexing local abundance and will provide a reliable means of monitoring recruitment and estimating recruitment variability. In contrast, standardised commercial catch rates in the core area have been highly variable throughout the history of fishing in the survey area and the age data do not show modal class progression suggesting they are not useful for monitoring recruitment. Stations in the experimental stratum near Ross Island had high catch rates and much larger fish than in the other strata, and warrant future monitoring due to their unique nature. We recommend the survey be continued to provide information on year class strength and an index of local abundance to be incorporated in the stock assessment.

In 2013, the Scientific Committee of CCAMLR could not achieve consensus on a stock status for Antarctic toothfish in the Amundsen Sea Region (ASR - Subarea 88.2 SSRUs 88.2C-H) and identified several areas for further work. This work was carried out and presented to WG-SAM in 2014, who noted that the assessments were unable to fit the patterns in the recaptures of tagged fish seen in SSRU 88.2H. As a consequence, WG-SAM requested further runs be carried out with emigration estimated in the model. However, the models that mimicked emigration in a single area assessment model failed to provide adequate fits to the observed tag recapture data (Mormede et al. 2014b). As an alternative approach, we present the preliminary development of a two-area stock assessment model.

A two-area model is a method for allowing explicit modelling of movement (migration) of fish into an assessment model, allowing the modelling of the populations in the southern ASR (88.2C-G) and the northern ASR (88.2H), as well as modelling the movement of fish between these two areas.

The inclusion of two-areas and migrations into the assessment model has both provided a more plausible explanation of the population structure as well as enabling a better account of the observed patterns in the tag recapture data. However, in this preliminary model some difficulties still remain. While the local biomass in the north and the migration to and from the northern ASR were able to be discerned, the biomass in the southern ASR and the associated proportion of that biomass migrating northward could not be resolved. This was because the parameters that represent the initial biomass in the southern ASR and the proportion of that biomass moving north were highly correlated, with little data to inform estimation of biomass in the southern ASR. As a result, estimates of the biomass and current stock status of the population from these models remains highly uncertain.
This paper is presented for consideration by CCAMLR and may contain unpublished data, analyses, and/or conclusions subject to change. Data in this paper shall not be cited or used for purposes other than the work of the CAMLR Commission, Scientific Committee or their subsidiary bodies without the permission of the originators and/or owners of the data.

Further development of these approaches is recommended, and will need to be further considered by WG-SAM in 2015. Furthermore, because reliable estimation of biomass in the southern ASR is currently not possible due to lack of data, we recommend that the acquisition of additional data to provide an index of abundance for the southern ASR be given a high priority.

Mark-recapture data for Antarctic toothfish from individual seamount features in SSRU 88.2H were analysed to estimate biomass trends on isolated seamounts. Biomass estimates were also calculated for SSRU 88.2H as a whole. The analyses indicate that:

• Fish seldom move among seamounts within the complex, and residence time on particular seamounts declines rapidly over 1–4 years.
• Fishing has occurred on almost every seamount in every year and usually in proportion to the level of tagging on the seamount in the previous year.
• Trends in local biomass estimates for individual seamounts and SSRU 88.2H overall showed a decline in biomass through time with a slight increase in biomass since 2012.
• The pattern in recapture rates of annual cohorts of tagged fish through time in SSRU 88.2H indicates a decrease in the percentage of the population tagged due to the annual immigration of untagged fish, along with catch and emigration of tagged and untagged fish resulting in a decreasing trend in biomass overall.
• Annual immigration also results in a progressive inflation of biomass estimates from mark-recapture data, and therefore biomass estimates are most accurate after 1 year at liberty, but still overestimated due to immigration.

This paper frames a discussion for improving the assessment of toothfish abundance for SSRUs 88.2C–G. We initially provide a characterisation of the fishery and a summary of available tagging and length frequency data. Although 880 tagged fish have been released in this region, only 2 tagged fish have been recaptured. It is likely that the lack of recaptures of tagged fish in this region has been caused primarily by the poor spatial overlap of released tagged fish with subsequent fishing effort.

By drawing on the success of tagging programmes in other CCAMLR fisheries, we develop an approach for improving the spatial overlap of the location of tagged fish and subsequent fishing effort. We identify four main grounds previously fished in SSRUs 88.2C–G. We recommend that in the short term (the next 2–3 years) it be made mandatory for vessels to complete all of their sets inside one or more of these fishing grounds as a condition of fishing in the slope region (SSRUs 88.2C–G). This condition could be relaxed in future years once sufficient tags have been recaptured to carry out an assessment of the stock in this region. We also recommend that the tagging rate in this region be increased to at least 3 tags per tonne in the short term. In addition to improving estimates of abundance a higher tagging rate and more recaptures will also increase information on fish movement within the Amundsen Sea slope which will help reduce uncertainty about the stock identity of toothfish caught in this area.

The aim of this paper is to review the management of the Ross Sea toothfish fishery and to identify key research objectives for the fishery over the next 3–5 years in relation to Article II of the Convention. The paper focuses primarily on Antarctic toothfish, as catches of Patagonian toothfish are negligible, and covers Subarea 88.1 and Subarea 88.2 SSRUs 88.2A and B. We begin by briefly summarising the management and operation of the fishery up to and including the 2012/13 fishing year. This includes the 3-year experiment from 2005/06 to 2007/08, the further development of the CCAMLR tagging programme and associated requirements, and other changes to the management of the fishery. We then identify uncertainties in our current knowledge that need to be addressed to fulfil the requirements of Article II. These include, for example, uncertainty in the biological parameters and stock assessment of Antarctic toothfish, uncertainty in its ecological relationships with predators and prey, and uncertainty over other ecosystem effects of fishing which can be addressed over the short to medium term. However, the need to further develop Management Strategy Evaluation and Management Procedures for the toothfish fishery in the medium-long term is also recognised. The purpose of this paper is to begin the discussion on medium-term research objectives for the Ross Sea fishery and the development of a medium-term research plan which could be adopted by the Scientific Committee.
 

We propose a multi-year and multi-member research plan using standardised longline gear to sample the toothfish populations in the northern areas (61° - 66° S) of SSRUs 88.2A–B. The purpose of the research as requested by the Scientific Committee (SC-CAMLR XXXII, paragraph 3.76) is to characterise the local toothfish populations found there to better understand stock structure, movement patterns and improve estimation of population characteristics by Ross Sea spatial population models. Additional outcomes of the research relate to mapping the bathymetry of the fishable area, documenting relative abundance of Patagonian and Antarctic toothfish, tagging toothfish for biomass estimation and for stock linkage studies, and collecting information on distribution, relative abundance, and life history of bycatch species.

This paper presents an analysis of data gained from an eight-year tagging programme of Patagonian toothfish (Dissostichus eleginoides) in Subarea 48.4. It describes the tagging procedure, the information gained about biology, growth and movement of Patagonian toothfish, and looks at the potential link between statistical subareas 48.3 and 48.4. The characterisation of tag recapture data from Subarea 48.4 shows that the tagging programme is successful in providing substantial information for the stock assessment.

Grenadiers (Macrourus spp.) are the main bycatch species in the exploratory longline fishery for toothfish in the Ross Sea. Previous studies concluded that acoustics methods could be used to index the relative abundance of grenadiers and be useful for exploring spatial distribution. Automated acoustic methods were developed to estimate grenadier distribution and abundance based on echo counting. These methods can be applied to large volumes of relatively low quality, opportunistically collected acoustic data. Trials using data from SSRU 88.1I showed positive correlations between acoustic targets and longline catches of grenadiers and toothfish. Single targets were most abundant within 250 m of the bottom at seabed depths of 700–1000 m, and revealed consistent spatial patterns, with higher numbers of targets on the eastern side of the Iselin Bank. The acoustic target strength distribution of single targets was similar to that predicted based on the expected size range of grenadiers. Variability in spatial coverage between years meant that it was not possible to obtain a consistent time-series of relative abundance estimates for grenadiers from acoustic data collected opportunistically by New Zealand vessels in SSRU 88.1I. The next step will be to apply to these methods to a wider set of data across the Ross Sea region.

Recommendation:
We recommend that other nations routinely collect acoustic data in the Ross Sea fishery and make this available for analysis. Collection of acoustic data from a larger pool of vessels would increase the available information, so that more consistent spatial coverage of grenadier distribution might be achieved.

Two species of grenadier are predominatly taken as bycatch in the Ross Sea region, Macrourus whitsoni and M. caml. A total of 220 otoliths from M. whitsoni and 307 otoliths of M. caml from fish taken as bycatch in the Ross Sea on New Zealand fishing vessels in the 2011–12 fishing year were used to test whether the otoliths of the two species can be discriminated. Samples of both species of macrourid were obtained from SSRUs 88.1B, 88.1C, 88.1H, 88.1J, 88.1K and from 88.2H. Lengths of M. whitsoni in the sample were 34.5–65.1 cm total length (TL), and 12.0–24.0 cm pre-anal length (PAL); lengths of M. caml in the sample were 34.5–81.5 cm TL and 11.0–30.0 cm PAL. Both males and females were included in the sample, but females predominated. A linear function of fish total length (cm), depth of the whole otolith (Depth, mm), and maximum cross-sectional area of the otolith (Area, mm²) gave excellent discrimination between the two species:

α = 1.254 + 0.03512*TL - 0.02463*Depth - 0.7668*Area

Where α<0.5 indicates M. caml and α>0.5 indicates M. whitsoni. Based on withholding data for testing in 10 folds, the multiple linear regression coefficient, R=0.776 [F(4, 522) = 272.9, p < 0.001], the proportion correctly identified was 92% (483 / 527), and the area under the receiver-operator characteristic (ROC) was 0.968. A similar discriminant function based on PAL (cm) rather than TL performed almost as well (90% correctly identified, ROC 0.97):

α = 1.330 + 0.09022*PAL - 0.02453*Depth - 0.6915*Area

There was no significant variation in the discriminant values (α) according to sex of the fish or for Subarea of capture for M. caml (CCAMLR Subareas 88.1 versus 88.2). However, there was a small but significant difference in α for M. whitsoni between subareas 88.1 and 88.2 [t(218) = 4.20, p < 0.001 ***], which provides some evidence that they could be separate stocks in the two CCAMLR Subareas.