CCAMLR

The behaviour patterns of Japanese krill fishery vessels in Area 48 were analysed using questionnaires on the reasons why the vessel changed their fishing grounds, which were sent out to of the Japanese fishing vessel since the 1989/90 fishing season. Among many reasons for changing fishing grounds, krill density, krill, size, ice condition, transshipment, and salp-by catch accounted for 95.6% of the changes. Although low krill density was the primary reason for changing fishing grounds, other seasonal factors such as greenness or ice condition could become important. A general picture of the seasonal succession of the Japanese krill fishing operation revealed that they tend to utilize fishing grounds close towards the southern limit within the ice free range. This pattern may well vary between nations, and it is essential to perform similar analyses for the other nation’s vessels. A conceptual model for Japanese krill fishing operation is proposed.

Growth trends of Antarctic krill with sex, length, season and region using over 10 years accumulation of instantaneous growth rate (IGR) measurements were modelled using a Linear Mixed Model (LMM). A model of inter-moult period (IMP) as a function of temperature, required to convert IGR to specific growth rate, was fitted to data from published constant-temperature rearing studies and this model was used to predict seasonal IMP using a model of the average sea-surface temperature seasonal trend for each region. Smaller krill exhibited higher growth rates and a progressive decrease in the IGR with increasing size was generally observed. This trend decreased from summer to autumn with small to negative values of IGR predominating across all size classes by autumn. The period of rapid growth was December in Indian sector, whereas in the Scotia Sea sector it appeared to be a few months earlier than this. Significantly lower growth rates were exhibited by females in January and February relative to males. Seasonal specific growth rates estimated in this study were compared to previous studies, and suggested that wild krill show more rapid growth over a shorter growth period than it was traditionally thought.

Patterns of fishing ground selection were characterized using STATLANT and CCAMLR fine scale data. Among the 15 SSMUs within Areas 48.1, 48.2 and 48.3 including the pelagic SSMUs, only one third of them were identified as the main contributors to the total catch. A shift of operational timing towards later months within fishing seasons was observed in Area 48.1 (Dec-Feb to Mar-May). However, operational timing stayed relatively constant in Areas 48.2 (Mar-May) and 48.3 (Jun-Aug). During a quarter century of krill operation in Area 48, patterns of SSMU usage has changed. Three different patterns of seasonal SSMU selection were characterized by following the result of cluster analysis. Frequently used SSMUs did not always match the areas of high krill densities observed by scientific surveys, and possible reasons for this mismatch were further discussed. Desire for revised data submission format was also recognized in order to accommodate any possible development of the fishing techniques.

Von Bertalanffy (VB) growth models for Antarctic krill have in the past been calibrated from population-level data consisting of modal lengths obtained from a time sequence of length frequency samples. We develop an alternative approach to predicting the trajectory of length over time using a step-growth function that combines models of instantaneous growth rate (IGR) at moult calibrated from direct measurements of individual pre- and post-moult krill sampled from the wild with a model of temperature-dependent intermoult period. Using summer and early autumn data for juveniles and males sampled from the Indian Ocean sector we model IGR as a function of pre-moult length and season using linear mixed models incorporating cubic smoothing splines. We generate a number of growth trajectories starting from an age 1+ mean length for different scenarios of winter and spring growth. We then provide convenient parametric approximations to these step trajectories using either punctuated-growth or seasonal-growth VB models. Our models indicate that, allowing for shrinkage, age 6+ mean length for the Indian Ocean sector was close to 53 mm compared to 57 mm obtained from studies for the Atlantic Ocean sector.

Krill carapaces measurements have been used to reconstruct krill length frequencies in Antarctic fur seal diet. The discriminant function currently used to determine sex, and the sex-specific allometric equations for calculating total length from carapace length, were derived from South Georgia krill populations. The equations have been applied to fur seal diet studies in the South Shetlands but until now have not been validated using locally sampled krill. This study reports on a three year study validating the use of discriminant functions to determine sex of krill based on carapace length and width and independently derives sex-specific regression models for krill collected in the South Shetlands. Allometric equations derived from South Georgia krill overestimated total length. Applying a discriminant function derived from mature krill in years following significant recruitment events with large proportions of immature krill resulted in significant bias towards male krill and an overestimation of krill length. We propose some standard guidelines for applying discriminant functions, allometric equations and for interpreting results.

A GIS tool to assist in the planning and designing of surveys of colonial breeding species is described. The tool could be applied to any region of Antarctica given available colony map and count data. The tool currently implements only very simple survey designs, but could be further developed to implement more complex and efficient designs. An example of its potential use is provided. In the example the tool is used to sample a population of Adelie penguins in the Holme Bay region of East Antarctica as might occur in a real a pilot survey, and the resulting sample densities used to predict the number of sampling units required to estimate abundance for the entire region with a pre-determined precision.

A summary of considerations and decisions made by the correspondence group on land-based predator abundance is provided for comment by the Working Group. In order to complete specifications for future survey work, the correspondence group seeks advice from the Working Group on the relevant spatial unit for estimation in Area 48 and the required precision for estimation in each unit. A workshop addressing survey design issues is proposed for 2006. Terms of reference for the workshop are provided for comment.

A survey was undertaken to provide estimates of the abundance of crabeater, Ross and leopard seal populations in 1,500,000 km2 of pack-ice off east Antarctica between longitudes 60-150oE. Sighting surveys were undertaken along almost 10,000 km of survey transect from an icebreaker and two helicopters to estimate the density of seals hauled out on the ice in survey strips. The probability of detecting seals in survey strips was estimated using double observer line transect methods, and satellite-linked dive recorders were deployed on a sample of seals to estimate the probability of seals being hauled out on the ice at any time of day. Due to non-random placement of survey transects, model-based inference, involving the fitting of a density surface as a function of geographic covariates, was used to extrapolate estimated densities in surveyed areas to the entire survey region. Estimating uncertainty in abundance estimates included consideration of uncertainties in species identification, estimation of detection probability, estimation of haulout probability, and extrapolation from sampled strips to the entire survey region.

Power analyses were carried out using a 12 year data set from the Béchervaise Island Adélie penguin colony with the aim of determining minimum sample sizes required to detect systematic temporal change in CEMP Parameter A5 (penguin foraging trip duration). Two different types of systematic change were investigated: 1) change occurring at a constant rate after a certain point in time and 2) a sudden step change in parameter values from one average level to another. Modelling showed that change of the latter form could be more quickly and powerfully detected at a range of effect sizes than could consistent rates of change occurring over longer periods of time. Sample sizes of 50 penguins carrying out three foraging trips each were required to detect a 35% step increase in foraging trip duration after three years or alternatively, with equal power, a 3.75% annual increase in trip duration after 12 years from the onset of change (55% overall increase). Inclusion of ice cover as a covariate in the analyses enhanced detection of change for birds monitored in the guard stage of chick rearing. However, caution must be exercised when incorporating covariates into power analyses, as inclusion of covariates to explain part of the natural interannual variability in a monitored parameter is only useful if the covariate itself is independent of the factors causing the systematic change.

We examine fledgling weights measured over 11 years at Béchervaise Island in relation to two assumed proxies of prey availability: breeding success and foraging trip duration. Concordance between the two proxies was apparent when considering guard stage foraging trip durations but this was not as strong for the crèche foraging trips later in the season. Fledgling weights which are measured at the end of the breeding season were more strongly correlated with later foraging trips than with earlier trips. We interpret these results as an indication of variable resources between the guard and crèche stages of the breeding season. In some seasons, there appeared to be constant levels of resources throughout the breeding season resulting in good breeding success with heavy fledglings or poor breeding success with light fledglings. In other seasons, there was a disparity between breeding success and fledgling weight. For example, low breeding success could occur in a season with heavy fledglings associated with long foraging trips during the guard period and relatively short trips during the crèche period. The concerns raised by Williams and Croxall (1990) that fledgling weight may increase with an associated truncation of the distribution in poor seasons for seabirds with prolonged chick-rearing periods is unfounded for the Béchervaise Island Adélie penguin population. It would be useful to determine the demographic consequences of variable fledgling weights in terms of subsequent chick survival for this population.