Validation-free estimation of chronological age via close-kin

SO_SCPLOWUMMARYC_SCPLOWAge is a fundamental life-history parameter in animal ecology and wildlife management. Age informs key ecological characteristics including population age structure, recruitment strength, extinction risk, reproductive maturity, and mortality rates. This importance has necessitated the development of chronological age estimation methods for wild animals. However, estimating chronological age is challenging for wild species, with noisy and potentially biased measures typically gathered from morphometrics, physical characteristics or, more recently, molecular methods like DNA methylation. These measures of age require at least some initial validation set of known-age individuals, or known time intervals, which is difficult to obtain for many species. Here, we present a solution to inferring the relationship between chronological age and error-prone observed age that does not require known-age individuals. The model couples the formulae for occurrence rates of half-sibling pairs, which decrease as a function of the birth-year gap between two sampled individuals, with time of capture. A pseudo-likelihood framework is developed for parameter estimation that can resolve linear and non-linear relationships and provide variance parameter estimates. We explore the methods efficacy for estimating chronological age using forward-in-time simulation and validate prior estimates of the relationship between vertebral band counts and chronological age for 3,000 school shark (Galeorhinus galeus) from an Australian fishery.