Allometry of natural mortality as a basis for assessing optimal release size in fish-stocking programmes

This study evaluates the use of general mortality-size relationships for the assessment of release size in stocked fisheries. Seven release experiments (53 stocking events) are analysed, using a survival model based on allometric mortality and linear-length growth, allowing variation between experiments in both the allometric exponent and the level of mortality at reference length or generalising in one or both of the parameters. Results support the existence of a consistent allometry that applies independently of the overall level of mortality. The best-performing model is one in which the length exponent of mortality is set to -1 a priori, while mortality at reference length is allowed to vary between experiments (ranging from 0.7 to 33 per year at 15 cm in the present study). Even though the allometry of mortality is constant, the relative survival advantage of stocking large fish increases with the level of mortality at reference length. Using the identified length exponent of mortality of -1, survival models are derived for the linear, exponential, and von Bertalanffy growth equations. The models can be used to assess alternative release sizes, given an estimate of mortality at reference length, to facilitate comparative studies and to aid in the design of release experiments.