Recapture probability underwater: predicting the detection of the threatened noble pen shell in seagrass meadows

Abstract

Capture‐recapture methods to estimate population size and its uncertainty rely on the important assumption that recapture probability does not change across individuals or within groups of individuals. This assumption is rarely met in natural populations and assessing the variance components associated with the recapture probability of marked individuals would help to design surveys. We illustrate a two‐session capture‐recapture survey of the threatened pen shell around the archipelago of Majorca, Spain. We were interested in investigating difference across teams of divers, sites, and individuals in the recapture probability of marked bivalves. We found that the probability of recapture was positively associated with shell width while differences across teams and sites were small. Our results indicate that inferences based on counts or on an average value of shell detection probability are underestimating total population size and in particular are skewed toward large size individuals leading to an erroneous measure of recruitment. In our case using an average value of detection probability, the number of young pen shells would be underestimated of c. 19%. Estimates of population size by capture‐recapture methods should be derived taking into account size‐dependent variability in recapture probabilities.