Models for Stock Assessment and Management

In many of the deterministic models widely used by fisheries agencies for stock assessment and management, it is difficult to take explicit account of the inherent uncertainties attached to key biological parameters. The models developed by CCAMLR attempt to incorporate some of the key effects of uncertainty into the analyses of various biological parameters and into the subsequent management advice being provided. To date, CCAMLR has taken a global lead in developing models of this kind to cope with specific uncertainties in stock assessments of both target and dependent species. Some examples include:

(i) In its first 8 to 10 years, CCAMLR adopted a single (i.e. fishery-targeted) species approach to managing fisheries. This was consistent with many of the conventional approaches to stock management of the time, even though these had many recognised weaknesses. CCAMLR’s efforts to manage multispecies fisheries and its pursuance of the ecosystem approach created a need for more realistic and complex models.

(ii) By 1994, CCAMLR had developed a ‘Krill Yield Model’ (KYM) to focus better its development of precautionary catch limits for the krill fishery. The KYM was developed in a manner which aimed to: (a) improve the meeting of the Convention’s objectives, (b) take more explicit account of uncertainty surrounding estimates of krill potential yield, and (c) applied clear, pre-agreed decisions rules to guide management decisions. An important output parameter of the KYM is g, which is intended to take into account variability in the life history characteristics (such as growth and mortality) of different krill stocks. This factor is then used in conjunction with a biomass estimate (B0) to obtain a precautionary catch limit consistent with pre-agreed decision rules. Such rules seek to maintain the spawning success of the stock concerned as well as its potential to meet predator food needs. The factor g is estimated in two stages and is currently fixed at 0.11. The output of the KYM will be refined as new data reduce the uncertainties associated with the estimation parameters. An important initiative in this regard aims to update the available estimates of B0, especially in the west Atlantic.

(iii) The ‘General Yield Model’ (GYM) follows a similar approach to the KYM and has been predominantly developed for finfish fisheries. Estimates of either current or pre-exploitation biomass, together with estimates of their uncertainties, are used. Recruitment fluctuations and uncertainty in biological parameters are also taken into account. The GYM enables CCAMLR to predict the effects of different levels of catch, even in the absence of direct estimates of abundance for entire stock(s). Precautionary catch limits can then be calculated.

(iv) The ‘Foraging–Fishery Model’ (FFM) attempts to formalise the description of functional relationships between krill and its predators. Since areas of highest krill-fishing activity are often close to the breeding locations of krill-eating birds and seals, broad areal estimates of krill biomass do not necessarily indicate the availability of krill to predators near the latter’s breeding colonies. Consequently, CCAMLR is developing the FFM to assess interactions, as well as the potential overlap (in both space and time) between the fishery and predators foraging for krill.

The multispecies models which CCAMLR is developing have no precedent. While in their simplified form the models do not require assessment of a large number of parameter values; the need to derive an estimate of the attached levels of uncertainty complicates their application considerably. Along with a growing database, such models have contributed greatly to CCAMLR’s development of a strategic and practical approach to describing the potential interactions between fisheries, harvested species and other dependent species.

Management Decision Rules

Management options are identified from the various model outputs. These are objectively selected to obtain catch limit(s) most congruent with the objectives of Article II of the Convention. As we have seen through application of the KYM, decision rules facilitate the taking of decisions in the setting, removing or varying of management measures taking into account assessments of harvested resource status, as well as the uncertainties attached thereto. Following the KYM, CCAMLR has also developed decision rules for the Patagonian toothfish fishery. It has also initiated strategic modelling aimed at setting scientific priorities in addition to developing and evaluating management options.