Accounting for tag-induced growth retardation in spiny lobsters using censoring and modelling approaches

Censoring (exclusion of data based on set criteria) and modelling approaches were used to estimate and compensate for perceived growth retardation of spiny lobsters Palinurus gilchristi tagged with internal anchor tags. Data were available for 2862 recaptured lobsters at large for periods of 0.01 to 14.66 years in five traditional fishing areas in southern South Africa. A generalized linear model indicated that growth increments were sex and area specific, with males growing faster than females. An individual-based model that simulated the effects of growth retardation on K and L∞ parameters was used to estimate the levels of censoring required to correct for increasingly severe effects. Longer growth retardation periods were easier to detect and correct for, and the undesirable effect of censoring in an unaffected population was minor. Censoring and modelling approaches provided similar estimates of growth parameters in most cases, although modelling often suffered from collinearity and apparent over-parameterization. Growth curves accounting for tag-induced retardation suggest that P. gilchristi grows substantially faster than previously thought, when no corrections were made. The effects of sample size on growth parameter estimates, and a reference point for management advice (F0.1), in the event of growth retardation, are shown.