An individual-based approach to evaluating experimental designs for estimating rates of fish movement from tag recoveries

Reliable estimates of fish movement rates from tag recoveries require an experimental design and a model for estimating quantities of interest from those recoveries. Many models have been developed, but experimental design problems have received little attention in the fisheries literature. This lack of a proper experimental design can severely compromise the accuracy and precision of the estimates of movement rates. In this paper, I develop a simple, individual-based, population dynamics model to account for fishing mortality, natural mortality, fish movement, tag shedding and differing reporting rates, derive a likelihood function for its parameter estimation, and use the population dynamics model and the estimation procedure to design a tagging program for the school shark Galeorhinus galeus. In the application, the minimum, mean, maximum, and three common norms of both relative bias and relative standard error of the estimates of all movement rates were each calculated as a function of the total number of fish released. From these calculations, one can readily determine the number of releases to achieve a certain level of precision and accuracy in the estimates of movement rates or vice versa.