CCAMLR

The recent fish survey at in the vicinity of Heard Island in 2000 has provided estimates of recruitment for at least one new cohort (1996) and another estimate of the 1995 cohort. The older cohorts are likely to be distributed more widely than the shallow part of the Heard Plateau, consistent with the observations from 1999. A reanalysis of all survey data from the region has shown that toothfish smaller than 450 mm are likely to be found predominantly in the shallow part of the Heard Plateau. Surveys aimed at providing a time series of recruitments could focus on this area. However, if estimates of mortality are to be obtained for this species then surveys over the deeper strata of the plateau will be required.

A recent 2000 survey of mackerel icefish in the vicinity of Heard Island enabled a revised estimate of yield for the coming 2 seasons. While the biomass of icefish appears lower than in previous years, it is dominated by a strong age 2 year class. Estimates of yield for the Heard Island Plateau population using the short-term assessment are 1150 tonnes for 2000/01 and 1000 tonnes for 2001/02 fishing seasons. The results of the 2000 survey confirm the need to retain a closure to fishing on Shell Bank. The paper reviews current catch controls and suggests that some consideration needs to be given to the implementation of closed periods during the year to provide safeguards to predators as well as to ensure spawning aggregations are not disrupted. Also, some consideration may need to be given to altering the minimum size regulations at CCAMLR in 2001 to account for the potential for unassessed cohorts to enter the fishery prior to assessments occurring.

A random stratified trawl survey was conducted in May 2000 to assess the abundance of icefish, Champsocephalus gunnari, and juvenile pre-recruit toothfish, Dissostichus eleginoides, on the Heard Island Plateau and Shell Bank. Abundances and length density plots of these species plus two principal bycatch species Lepidonotothen squamifrons and Channichthys rhinoceratus are given for each stratum. A total of seven surveys have been conducted in the Heard Island and McDonald Islands region since 1990 and results from each survey are presented so that inter-annual comparisons can be made. Abundance of D. eleginoides, L. squamifrons and C. rhinoceratus in each stratum are relatively steady, varying by a factor of 2-3 between years, but C. gunnari abundance can vary by a factor of about 20 between years. This is the result of occasional strong year classes entering the fishery and dominating the population for about 2 years before disappearing, as has been documented in other locations.

For the last 3 years, WG-FSA has accorded high priority to development of a method to integrate different indicators of stock status into assessments using the generalised yield model (GYM). In this paper, we propose a method, based on use of the sampling/importance-resampling algorithm, for incorporating information on trends in standardised CPUEs into GYM assessments. The use of the method is illustrated using data for D. eleginoides in Subarea 48.3. As the method requires only very small adjustments to the computer program implementing the GYM assessments, we propose that this method should be adopted for use during the 2000 meeting of WG-FSA.

Fish Heaven is a simple spatially explicit age structured fish model, containing the basic features of standard fish modeling with a number of optional extras. It was designed within an object oriented framework to allow for the possible extension in a number of directions. In particular it was designed with the view of testing the efficacy of different statistical sampling regimes given various plausible spatial structures of fish stock.
The model has a simple age and space structured population governed by the normal life history parameters and subject to both environmental stochasticity (if desired) as well as demographic stochasticity. It is a spatially explicit model in which fish will tend to distribute themselves according to their density and the underlying habitat quality variable. On top of this, fishing occurs. This includes a wide range of fishing options. The software has been constructed to allow for it to be extended to include more complex fishing management practices.
The manual for the software, Fish Heaven Version 1.0.0, forms the body of the paper. Although Version 1.0.0 of the software is complete it is still seen as a piece of software in development. Currently the software is designed to allow environmental simulations with fishing while capturing various statistics about the status of the system. Analysis of the output must be done with a separate piece of software, such as a statistical package, reading in the output files of Fish Heaven. A number of extensions are planned for the software and these are described in the appropriate sections of the manual as 'further improvements'.
The manual and software are both contained in the file FishHeav.zip. This is on the CD of Australian Antarctic Division software and additional copies are available from the Australian Antarctic Division. The manual contains colour pictures, although black and white copies of the manual still contain all of the necessary information. The copy of the manual in FishHeav.zip is a pdf file and includes all the coloured pictures.
The executable has been compiled for Windows NT but should run on any 32bit Windows operating system. It requires no additional libraries to run and takes up less that 1.4M of disk space. When running the size of the spatial grid is the limiting factor on memory requirements but is not terribly restrictive.

A brief description of hawl-by-hawl data sets from Ukrainian commercial vessels fishing tooth-fish in Subdivision 48.3 during 1986-199 1 is presented. The fishery was mainly carried out in summer months off the eastern South Georgia and at Shag Rocks. The length frequency modes amounted to 95-100 cm, which exceeded modes for the latest years of fishery. Using the generalized additive model the catch per effort estimates for 1986-99 were standardize. The fishery efficiency in 1986-95 was approximately similar, while in 1996-99 it decreased sharply. It is shown that significance of such factors as a month and depth is low, and a modified model developed without these factors consideration provides more accurate estimates.