New methods to derive sea turtle nester abundance from nest counts: Ground truthing and the bias of current approaches

Abstract

Abundance is a key parameter for characterizing animal populations of conservation concern. In sea turtles, annual clutch counts represent the most common population index. When the number of annual nesting females (ANF) is sought, clutch frequency (CF; average number of clutches laid per female) is used as a conversion factor. Here we assess the performance of five current methods for estimating ANF and of two new methods based on the interdependence of nesting activity parameters, investigated through a Monte Carlo approach with generalized additive models. All methods were evaluated on virtual populations and on a unique case of empirical data from a sea turtle population with a known ANF based on genetic identification of clutches (Caretta caretta, Northern Recovery Unit, USA). Results show that current methods produce errors often greater than the simple count of nesting females directly encountered, except for a capture-mark-recapture (CMR) model. One of the new methods takes advantage of both nest count and CMR data and shows the best performance at a wide range of detection probabilities. The second new method is based on nest count data only and can provide at least an approximate estimate of ANF where no CMR data are available. Our findings can dramatically improve sea turtle population abundance estimates worldwide. We recommend abandoning the use of any CF estimate as a direct conversion factor in favor of implementing the CMR model and the new methods to improve current estimates and to provide estimates where limited data have prevented doing so previously.