Sea turtle populations are overestimated worldwide from remigration intervals: correction for bias

Abstract

Estimating population abundance is key for species of conservation concern. This is particularly challenging for marine animals, like sea turtles, with ocean-scale distribution and migratory nature. However, sea turtles lay clutches on land where they can be easily counted; thus, clutch number has always been the most common index of population abundance. A female typically lays >1 clutch per year and does not reproduce every year. Therefore, 2 conversion factors are needed to convert the number of egg clutches to the number of adult females: the number of clutches laid by a female in a nesting season and the fraction of adult females reproducing in a season, which is linked to the breeding periodicity. The effects of breeding periodicity, probability of detection and annual survival probability on the derived adult female abundance were investigated by simulating a virtual population of adult females over a 15 yr beach monitoring period. The results indicate that current methods may greatly overestimate the abundance of sea turtle populations, especially in situations with a low detection probability, including temporary emigration. The factors involved and ways to minimize biases and errors are discussed, including a method which is easy to implement using existing datasets. A careful reassessment of current estimates of sea turtle abundance derived from nest counts and capture-mark-recapture data would be appropriate, and the potential error associated with such estimates should be considered when they are used in conservation status assessments.