CCAMLR

Trends in mean length and maturity stage of toothfish were examined using observer data from the 1998, 1997 and 1996 toothfish seasons. Toothfish are distributed down the shelf slope in relation to their size, an approximately linear relationship existing between depth and fish length. A statistically significant pattern of changing mean length at depth with month was found in both 1997 and 1998 seasons, suggesting movement of animals both up and down slope at particular times of year. Months with a high mean length at depth are May and August, and with a low mean length at depth are April and July. Examination of maturity data suggests that in addition to a major spawning event in late July/August, there may be a small spawning event in April/May. The observed shifts in length at depth may be associated with movement of animals to preferred spawning depths. A general model would be that toothfish form spawning groupings at about 1000-1300m depth. Just prior to spawning, animals move (generally downslope) to this depth, dispersing immediately afterwards, mostly upslope. Behaviour appears to be different at Shag Rocks and South Georgia, but there is some evidence for exchange of animals immediately post spawning between the two areas. There may be two (or more) spawning periods during the winter, with the larger spawning period being late July and August.

A monitoring program of demersal fish in inshore sites of the South Shetland Islands has continued in Potter Cove from 1991 to 1998, covering a continuous sampling period of 15 years and in Harmony Cove, Nelson Island, in the austral summer 1995/96. The decline in trammel net catches of fjord fishes of the species Notothenia rossii and Gobionotothen gibberifrons in relation to the non commercially fished Notothenia coriiceps, which was already reported for the period 1983-1990 in a previous study, is still evident. These results are supported by our knowledge on the diet of the piscivorous Antarctic shag Phalacrocorax bransfieldensis in the South Shetland/Antarctic Peninsula area in this decade. The most likely explanation for the decrease in recruitment to the inshore sub-populations of N. rossii and G. gibberifrons in the last 15 years is the effect of the offshore commercial fishery in the area in the late 1970s. This interpretation is consistent with the information on the historical offshore commercial fishing and with the results of scientific surveys in the area.

A random stratified trawl survey of Mackerel icefish in 2 populations (Plateau and Shell Bank) in the vicinity of Heard Island enabled a revised estimate of yield for the coming 2 seasons in 1999 and 2000. While the abundances of icefish are lower in 1998 than in 1997, the Age 3 and Age 4 cohorts are still sufficiently strong to provide for a yield greater than estimates of the long term precautionary yield. Estimates of yield for the Heard Island Plateau population are 732 tonnes for 1999 and 518 tonnes for 2000. These figures may need to be re-evaluated in light of known catches in the 1998 season since the survey. Estimates of yield were not made for Shell Bank given the very low abundance in this population. Unlike the previous 3 years, the Age 2 cohort in 1998 is very weak and is expected to contribute little to the biomass in subsequent years. If recruitment to Age 2 in 1999 is also weak then the fishery in 2000 will be predominantly on Age 5 fish. After this time, catch limits may need to be set at an estimate of a long term precautionary level and be maintained thereafter unless a further survey again showed that abundant cohorts had been recruited.

During the 1 997198 season, one vessel, Chilean registered stern trawler Betanzos, fished commercially for mackerel icefish at South Georgia (Subarea 48.3) using a midwater trawl for ten days between 25 December 1997 and 5 January 1998. The total catch of C. gunnari was 5.04 tonnes in 34 hauls. 67% of this catch was taken in just two hauls, confirming the patchy distribution of this species. Most of the catch was made up of fish between 22 and 30 cm long. Fish of this size range have previously been shown to be ages two and three. The catch of species other than C. gunnari amounted to only 0.21 tonnes. The vessel's fishing master had no experience of fishing for icefish and was not well briefed before the short trip to South Georgia. It is unclear whether the poor catches by the FV Betanzos were due to a low standing stock of the target species, or the inexperience of the fishing master.

A prerequisite for validation of an aging technique is a clear set of consistent aging criteria to test. Two sets of criteria ate presented for aging the otoliths of Dissostichus eleginoides, one (C1) developed from previous work on the same species; the other (C2) adapted from generalized criteria used to age high latitude fish from the north-west Pacific Ocean. Growth structures were found to occur at several scales and vary both within and between otoliths. Pairwise comparisons of repeat readings by one observer and readings by two observers showed considerable inconsistency in aging both within and between observers, reflecting the complexity of the otolith structure. Aging using criteria C1 was more precise than aging using C2. but tended to give higher age counts in younger fish than over the equivalent region in older fish. Both aging techniques indicated D. eleginoides to be a long-lived fish, but C2 indicated that captured fish may be mostly young, between 3-9 years old. The low precision for aging D. eleginoides indicates that techniques to maintain aging consistency without bias and within standardized limits of variation are likely to be important.

Longline sink rates were investigated using Time Depth Recorders on a bottom autoline vessel F.V. San Aotea in New Zealand. The objective of the project was to determine line sink rate, and the effect that adding weights to the line had on its sink rate. The vessel used Mustad autoline equipment that is designed to sink without weights, so non-weighted longline line sink rate data were collected initially to give an information baseline. Further trials were then conducted using added weights as would be used in normal fishing operations to test the effectiveness of weighting the longline as a method of accelerating line sink rate and thus avoiding incidental capture of seabirds. A new rapid attachment method for Time Depth Recorders was also developed and is documented. The study found the middle of an unweighted longline of this design sinks to 10m in a mean time of 63.0 seconds (n = 11, c.v. 16.7%), compared with the start of the longline which takes a mean time 31.1 seconds (n = 11, c.v. 30.4%) to reach 10m. The tori line aerial section covered the longline for a mean time of 26.3 seconds (n = 25, c.v.13.6%). The longline weighting trials indicate that the weighting regime used had no detectable effect on the overall line sink rate. However, observation indicated that the weighting regime did have quite noticeable effects on line sink rate for 20 - 40 m either side of the attached weight. Given the data collected on line sink times and tori line coverage it would seem that we need quicker sink rates to substantially decrease the incidental mortality of seabirds during auto line fishing.