CCAMLR

We measured sink rates of longlines using two different methods: one was string wrapped around bottles and the other electronic time-depth recorders (TDRs). Bottles provided faster sink rates TDRs. Results obtained from the bottle method were more variable, and hence less reliable, among individual deployments than TDRs requiring repeated deployments on a longline. Sink rates measured to 2 m with bottles averaged 0.13 ± 0.02 m.s-1 compared to 0.21 ± 0.07 m.s-1 recorded by TDRs. This difference was statistically significant (t5 = -2.720, P = 0.042). Measuring sink rates to > 10 m proved difficult and unreliable with bottles. Bottles had a higher failure rate (51 – 68 %) than TDRs (> 90 %), for example, with string becoming entangled or bottles vanishing from sight. We recommend while bottles may be an appropriate means for measuring sink rates to shallow depths (> 5 m) that they should not be used for estimating sink rates to greater depths.

The summer diet of two species of icefishes (Channichthyidae) from the South Shetland Islands and Elephant Island, Champsocephalus gunnari and Chaenocephalus aceratus, was investigated from 2001 to 2003. C. gunnari fed almost exclusively on krill (Euphausia superba) in all years. The importance of other taxa (Themisto gaudichaudii, mysids, myctophids) in the diet was negligible. The average feeding rate of C. gunnari inferred from an exponential gastric evacuation model was between 1.0 and 1.5% body weight per day. Most of the stomachs of C. aceratus were empty. Stomachs with food contained mainly krill, mysids and fish. Among the fish taken, locally abundant species formed the bulk of the diet. They were Gobionotothen gibberifrons in 2001, Lepidonotothen larseni and C. gunnari in 2002 and L. larseni in 2003. An ontogenetic shift in feeding preference of C. aceratus was observed: Fish smaller than 30 cm fed on krill and mysids, while larger animals relied primarily on fish.

We studied the food and feeding of five species of icefish in the Elephant Island – South Shetland Islands region in March-early April 2003. C. gunnari took primarily krill. C. aceratus fed on krill and fish with some differences found between Elephant Island and the South Shetland Islands. The bulk of the diet of C. rastrospinosus consisted of krill and fish. Diet intensity, however, was low due to the progressing spawning season. C. antarcticus took almost entirely fish while P. georgianus was a fish feeder at Elephant Island and preyed on krill and fish further south. A large proportion of the stomachs was found to be empty in C. aceratus and C. antarcticus.

The 2002/2003 fishing season was the first in which three new policies for rays were tried:
• It was the first in which vessels were asked to cut off all rays at water level to assist survival and conservation of rays;
• It was the first in which very good quality information was available from all observers’ tally periods on the catch rates of rays cut off the line;
• In 2003 the UK, with the cooperation of Chile and one fishing company, undertook an experiment to determine the survivorship of cut-off rays. This is described elsewhere by Endicott & Agnew (WG-FSA-03/??).
We estimated the number of rays that were likely to have died as a result of being caught on longlines using a combination of the observer tally information and the survivorship experiments. The estimate, between 34 and 46 t, is an order of magnitude lower than the precautionary TAC for this fishery, 390 t.

Skate post haulage survival experiments were conducted by one Chilean fishing vessel throughout the months of May and June, during 2003 toothfish fishing season in Subarea 48.3. From these results we can conclude that depth has a significant influence on the mortality of skate, with a predicted mortality of 0% for skate caught in waters of less than 1100m. A predicted mortality of 65% for skate caught in waters between, 1400 to 1600m and a mortality of 100% for skate caught at 2000m. We recommend from these findings that further survival experiment be conducted, particularly in more shallow waters where our sampling was restricted. Our findings support the suggestion of cutting ray from the line before landing (FSA-02 Para 5.78), as this will help ensure survival chances of skate are maximised. Results from this experiment can also be used to calculate the expected skate mortality within the fishery, based on depth of capture.

Estimates of bird catches by IUU fishing made using the new method presented by Agnew and Kirkwood last year (WG-FSA-02/4 & 5) are updated to take account of the comments of the working group. Last year we suggested that the Isabel data from 1997 were rather atypical of bird catch rates in the illegal fishery, and that these were more likely to be similar to the bird catch rates made by other vessels in 1997. Including the Isabel data and weighting the bootstrap estimates of bird catch rate by the number of hooks observed leads to revised estimates of total catch in 1999, 2000 and 2001 of 1115, 4898 and 1236 birds respectively. The relevant rates per 1000 hooks for this bootstrapping option, which should be used by WG-FSA/IMAF in future estimates of bird bycatch rates are:
a. Summer: Median: 0.741 birds/1000 hooks; 95% confidence limits – lower: 0.39, upper: 11.641
b. Winter: Median: 0.0 birds/1000 hooks; 95% confidence limits – lower: 0.0, upper: 0.99.

Reasons of differences in icefish distribution and biomass assessment during two surveys made by STM “Atlantida” are revealed. Availability of large icefish concentrations in 2000 in the northwest shelf area in 10-100m layer resulted in formation of adult fish aggregations in water column. The year 2002 was characterized by krill distribution in near-bottom layer, causing icefish presence near the bottom even in dark hours of the day. During the two surveys, fry and immature fish occurred in numbers in pelagic zone. At present, the population mainly consists of immature and first maturing fish, which stay in water column for considerable part of the day. Earlier beginning of pre-spawning period, as distinct from mean annual, was noted in 2002 in the northeastern part of the area.
It is shown that pelagic part of the stock permanently existing during the day is not taken into account by bottom net survey, which, consequently, is disregarded when estimating the TAC. Diverse size groups, mainly juvenile and immature fish, form pelagic and near-bottom components of the stock. Under certain oceanological conditions, length-age composition of pelagic component may be formed by all size classes.
Results of the fishery and investigations in Ch. gunnari habitat show that methodology of census surveys does not conform to their distribution pattern. Conduction of net and acoustic survey will enable a quantitative assessment to be made for bottom (bottom net survey) and pelagic (hydroacoustic survey) components of Ch. gunnari stock. Taking Ch. gunnari pelagic component into account will allow computing the recruitment that will enter the fishery the year to follow. A need to change the survey season to avoid underestimation of fish biomass related to its distribution is demonstrated.