Bias in the estimation of krill yield from using a deterministic formulation of the median unexploited spawning biomass
Two Monte Carlo simulation models, the krill yield model and the generalised yield model (GYM), are currently available to CCAMLR that can be used for estimating yields according to decision rules that relate spawning stock status to the median unexploited spawning biomass. These models are used in order to account for uncertainty in the biological parameters and abundance of the stock being assessed. CCAMLR currently uses results from the krill yield model as a basis for setting catch limits in the krill fishery, while the GYM is used for assessing yield in other finfish fisheries but can also be used for krill. The decision rules relate to the effects of a specified long-term annual yield on (i) the probability of the spawning stock being reduced below a set proportion of the median pre-exploitation spawning biomass during a projection trial and (ii) the median status of the spawning stock at the end of the projection trial (as a proportion of the median pre-exploitation spawning biomass). These rules require that the median pre-exploitation spawning biomass is known prior to a projection so that in each trial a direct assessment can be made of (i) whether the stock became critically depleted in that trial and (ii) the status of the stock at the end of the trial. These values from all trials are used to determine the probability of critical depletion and the median level of escapement respectively. The Monte Carlo simulations take account of uncertainty in the biological parameters used to estimate yield. In this respect, the biological parameters may change between individual projection trials. Consequently, the median pre-exploitation biomass needs to be estimated prior to each trial in order for it to be based on the parameters for that trial. The krill yield model uses a biased estimate of the median pre-exploitation spawning biomass with a method for correcting this bias in the level of escapement. The GYM uses an unbiased estimate of the median pre-exploitation spawning biomass and, thus, requires no correction. This paper examines the effects of the biased estimates of the median pre-exploitation spawning biomass on the estimates of krill yield. The results show that the biased estimate results in a bias of the probability of critical depletion, while the level of escapement is not appreciably sensitive to the method of estimating the median pre-exploitation spawning biomass. This bias in the krill yield model will result in the catch level given for a set probability of critical depletion being too high. However, this will not affect the estimates of long-term precautionary yield currently adopted by CCAMLR for the krill fishery because the level of escapement is the binding rule in this case. The results indicate that the GYM using the unbiased estimator of the pre-exploitation biomass is the preferable model for assessing long-term annual yields.