CCAMLR

A population dynamics model for Adelie penguins is developed which takes account of the factors of lower adult survival rates in the year of first breeding, and the possibility that such breeding may be deferred one year if conditions are poor. A Bayesian approach for estimation of the model parameters is put forward; its estimation performance (including robustness to certain mis-specifications) is tested by simulation, and it is applied to preliminary data for penguin colonies in the South Shetlands for illustrative purposes. The results for this case would likely be improved by the provision of more comprehensive data. Extensions of the overall approach are suggested, but a prerequisite for their pursuit is discussions on the likely levels of measurement error in the observational data.

A method is put forward which estimates the proportions by which the annual consumption of krill is split between predator species. This method uses data on the length distribution of krill from predator stomachs. The proportions estimated depend, through the krill yield model, on the schedule of krill natural mortality at age. In principle, such a method would allow inferences about this schedule to be drawn by comparing the proportions which it predicts to independent estimates based on predator abundance and food consumption studies. Some results of the method are put forward purely for illustrative purposes. Further discussions about the practicalities of obtaining reliable estimates of the inputs required by such an approach are required before it is taken further.

The krill-predators model is generalised by splitting post-first-year animals into adults and sub-adults, with different survival rates for these two groups. Broadly speaking, the sub-adult survival rate can be adjusted to ensure that the modelled predator population reflects observed growth rates, thus solving problems of lack of self-sustainability encountered in earlier model analyses. This solution brings with it new problems, however, in that two further parameters need to be specified: one reflects the difference between adult and sub-adult survival rates, while the other relates to the split of the overall density dependence effect between the first year and following sub-adult years of life. Since evaluations of the level of resilience of a predator to krill fishing depend on the values for both these parameters, consideration needs to be given to the possibilities for qualitatively inferring these values or estimating them from data. The generalised model is applied, for illustrative purposes only, to the Bird Island blackbrowed albatross population. However, even as generalised, the model is unable to achieve compatibility with survival rate data for the Antarctic fur seal coupled to an annual population growth rate of 10%.