Exploratory longline fisheries targeting mainly Antarctic toothfish (Dissostichus mawsoni) have been operating in Divisions 58.4.1 and 58.4.2 since 2003, however robust stock assessments and catch limits according to CCAMLR decision rules have yet to be developed. A key step in developing management advice in these Divisions is understanding the distribution, abundance and boundaries of the target stock. WG-FSA-16/29 outlined the first multi-member toothfish exploratory fishery research plan, including research objectives, methods and milestones. Within this plan, milestones 1.5 and 2.1 involve the development of stock structure hypotheses and spatially-explicit habitat-use models. Accordingly, this paper reports on the development of generalised additive mixed models to characterise relationships between environmental variables and (1) relative density, (2) mean weight, (3) proportion mature and (4) sex ratio of D. mawsoni, including in adjacent Division 58.4.3b. Spatial heterogeneity in catch composition indicated that D. mawsoni were not randomly distributed across the area. Models were used to generate predictions of D. mawsoni density and composition across a broad spatial scale, and revise hypotheses relating to the structure and functioning of the stock.

This paper presents an updated assessment for the Patagonian toothfish (Dissostichus eleginoides) fishery at Heard Island and McDonald Islands in CCAMLR Division 58.5.2 with data until end of the season 2015/16. The updated assessment model is based on the best available estimates of model parameters, the use of abundance estimates from a random stratified trawl survey (RSTS), longline tag-release data from 2012-2015 and longline tag-recapture data from 2013-2016, and auxiliary commercial composition data to aid with the estimation of year class strength and selectivity functions of the trawl, longline and trap sub-fisheries.

Compared to the 2015 assessment that was accepted by WG-FSA-15 to be used for management advice, this assessment takes into account (1) new fishery observations up to the end of 2016 including new ageing data from the RSTS and commercial fishery from 2015-2016, (2) updated growth parameters, (3) updated maturity parameters, (4) updated tag loss estimates (5) a bias correction for fish emigrating out of Division 58.5.2, (6) survey biomass and catch proportions instead of survey abundance numbers, and (7) iterative data weighting using the ‘Francis’ method. All model runs were conducted with the CASAL version 2.30-2012-03-21 rev. 4648 that was agreed on by WG-SAM-14.

The updated assessment model leads to a smaller estimate of the virgin spawning stock biomass B₀ than that obtained in 2015, with an MCMC estimate of 77 286 tonnes (95% CI: 71 492-84 210 tonnes). Estimated SSB status was 0.61 (95% CI: 0.58-0.64). Despite the smaller biomass, changes to the model compared to 2015 in particular its higher productivity with the updated maturity parameters, meant that the catch limit that satisfies the CCAMLR decision rules has increased from 3405 tonnes to 3525 tonnes.

CCAMLR XXXIV (para. 5.68) endorsed new pre-season (1-14 April) and post-season (15-30 November) trial season extensions, in addition to existing pre-season (15-30 April) and post-season (1-14 November) trial extensions in the Patagonian toothfish longline fishery in Statistical Division 58.5.2 (WG-FSA-15/48).

Australia undertook to report annually on the results of all the trials. This paper presents information on fishing effort and seabird interactions with fishing gear collected during the periods 1-14 November 2016, 15-30 November 2016, 1-14 April 2017 and 15-30 April 2017. This report complements the fishing effort and interaction data reported in 2015 (WG-FSA-15/48) and 2016 (WG-FSA-16/28r1).

Tag-loss or tag-shedding rates for tagged fish in the Patagonian toothfish (Dissostichus eleginoides) fisheries at Heard Island and McDonald Islands (HIMI) in Division 58.5.2 were estimated following the method proposed by Adam & Kirkwood (2001). Tag-loss rates were generally low, with longline-caught and recaptured fish losing their tags quicker than trawl-caught and recaptured fish. With an initial tag-loss proportion of 2.8% and ongoing tag-loss rate of 0.056 in the longline fishery, 99.3% of all double-tagged fish retain at least one tag after one year, and 80.2% after ten years. The parameter of annual tag-loss rate in CASAL’s single-tag model was approximated, for a maximum time at liberty of 6 years, as l = 0.014 for longline and l = 0.007 for trawl. However, this single tag-loss rate for longline varied strongly between time periods, with l = 0.007 for 2003-2006, l = 0.021 for 2007-2011, and l = 0.006 for 2012-2015. We recommend that for longline releases the parameter estimates for these time periods be used in the stock assessment for Patagonian toothfish in Division 58.5.2.

This study investigated maturation and spawning of Patagonian toothfish Dissostichus eleginoides in the Heard Island and McDonald Islands (HIMI) fishery on the Kerguelen Plateau in the Indian Sector of the Southern Ocean based on gonads and otoliths collected between years 2004–2015, and using histological analyses and calibration of macroscopic staging criteria. Dissostichus eleginoides at HIMI spawn throughout the late autumn and winter months of May–August, and spawning activity is concentrated on slopes along the west and south of the plateau around HIMI at depths of 1500–1900m. Comparison between histological analyses and macroscopic gonad staging indicated that many fish that had spawned, as indicated by the presence of post-ovulatory follicles, returned to a resting stage which was macroscopically indistinguishable from maturing fish. Furthermore, the occurrence of females of all size classes with low gonadosomatic index and low macroscopic gonad stage during the spawning season suggested that a proportion of mature females did not spawn every year. Age-at-maturity estimates, obtained based on the assumption that fish of macroscopic stages ≥ 2 were mature, decreased between the 2004–2009 and 2010–2015 periods for both sexes. However, the magnitude of this temporal variation in age-at-maturity varied between gear types and fishing depths, and variable sampling regimes likely influenced these variations. This study highlights the importance of correct interpretation of macroscopic gonad stages, and understanding the influence of fishery operations on estimations of life-history parameters.

The annual random stratified trawl survey was undertaken in Division 58.5.2 in the vicinity of Heard Island and the McDonald Islands during April 2017. This paper provides an updated growth model and a preliminary assessment for mackerel icefish (Champsocephalus gunnari) population structure, abundance and yield in Division 58.5.2 to the west of 79^o 20’ E using standard CCAMLR methods (CMIX and Generalized Yield Model).

The 2017 survey showed a large 3+ cohort dominating the population. An updated growth model based on recent cohort progressions indicated an increase in size-at-age compared to the estimates of Welsford (2010), however Generalized Yield Model projections indicated that the updated growth parameters resulted in only a small impact on yield.  Catches of 526 t in the 2017/18 season and 395 t in the 2018/19 season respectively satisfied the CCAMLR decision rules. Given their apparent plasticity in growth, we recommend that growth parameters used in assessments be updated routinely.

This paper includes updated local biomass estimates of toothfish populations with bootstrapped confidence intervals using both CPUE-by-seabed area and Chapman biomass estimation methods agreed at WG-SAM-16 with revisions to the CPUE-by-seabed area parameter values agreed at WG-SAM-17. The analysis includes Research Blocks in Subareas 48.2, 48.6, 58.4 and 88.3. Given the changes in CPUE-by-seabed area method parameter values that were agreed at WG-SAM-17, the estimates of biomass in the relevant Research Blocks have increased relative to the estimates that were presented at WG-FSA-16. The higher CPUE-by-seabed area estimates has also resulted in a greater overlap in confidence intervals with the Chapman estimates in some, but not all, of the Antarctic toothfish Research Blocks. 

Fatty acid (FA) profile and stable isotope (SI) signature analyses has been used to examine feeding ecology of Antarctic toothfish and food webs of Antarctic marine ecosystem. Specimens of Antarctic toothfish, by-catch and their stomach contents were collected for the past five years in the eight locations of he Antarctic Waters. We investigated ontogenetic and spatial variability in resource utilization of Antarctic toothfish and distribution patterns of FA profiles in various organisms because of the particular importance of accumulation of lipid deposits in their tissue. Toothfish trophic niche indices was calculated using both FA profiles and SI values. Results shows spatial variation in their resource utilization between the Ross Sea shelf (RSS) and the deep-sea slopes the western Indian Ocean sector (IOS) and the slope towards the Pacific Ocean sector (POS). FA profiles of toothfish were characterized by high proportions of monounsaturated FAs and low proportions of polyunsaturated FAs, differing from other organisms in their habitats. Distribution patterns of FA profiles among various organisms were well established by self-organizing map (SOM) analysis. Similar FA profiles and SI values of toothfish between ISO and POS indicate a depth-dependent dietary difference from that of RSS. This appears to be closely associated with ontogenetic and spatial variation in resource utilization during their migration. The combination of both tracers explained well their dietary utilization pattern. Indeed, a recently developed mixing model calculation for FA profiles and SI values, which follows the Bayesian machiner, highlights the nutritional importance of Noththenidae fish compared to Channichthyidae species to whole population level. In contrast, the model calculations provided an estimation of dietary shift during ontogeny and thereby migration to deep water. Finally, the model results clearly explained spatial shifts on the nonmetric multidimensional scaling plot and the SOM map by realistic diet proportions.

Japan and France revised the next season’s (2017/18) research plan in research blocks 58.4.4b_1 and 58.4.4b_2 using the latest CCAMLR C2 and Observer data, following several advices during WG-SAM meeting in 2017. We have used extracted data provided by the CCAMLR Secretariat on 27th July, 2017.

The estimated median stock size in block 58.4.4b_1 and 58.4.4b_2 was 1452.03 and 921.91, tonnes respectively, in Chapman method considering tags released in the last three years as effective for the biomass estimation. The estimated median stock size in block 58.4.4b_1 and 58.4.4b_2 was 484.51 and 498.76 tonnes, respectively, in CPUE analogy method (reference area: HIMI) using the relevant information of reference area (recent median CPUE, estimated biomass and updated area size) recommended in the WG-SAM-16/18 Rev.1.

We propose to continue the current research operation for the next fishing season with the same survey design and total catch limit of 70 tonnes to further strengthen the stock assessments in the area.