Use of Bayesian hierarchical models to estimate northern abalone, Haliotis kamtschatkana, growth parameters from tag-recapture data

Bayesian hierarchical models were developed to estimate the growth parameters of northern abalone, Haliotis kamtschatkana, using tag-recapture data with a mixture of single and multiple recaptures. Individual variability in the growth parameters L∞ and k of the von Bertalanffy model was incorporated in the analyses. The models developed fit the data well based on the Bayesian p-values. Variability in L∞ for individuals was high relative to the variability in L∞ for the population, and variability in k for individuals was about the same as the variability in k for the population. Simulations showed that estimates of the growth parameters were accurate (relative biases <5%), when variability in both L∞ and k or just in L∞ was accounted for. The “true” values of the parameters, L∞ and k, were contained in the estimated 95% credibility intervals in 90–94 out of 100 simulation runs on 100 simulated data sets. Overall, allowing for variability for both L∞ and k resulted in moderately more accurate estimates than allowing for just L∞. On the contrary, estimates were unreliable when variability in just k was considered. Using the WinBUGS software program, the calculation procedure was rather simple irrespective of which growth parameter was modeled with variability.