Mixed models, statistical power and sea turtle nesting beach surveys

Full Title: Mixed models, statistical power and sea turtle nesting beach surveys. How long and when?

Participants: R. Bjorkland (lead)¹, M. Sims¹, J.I. Richardson² and P. Mason³

1-Duke University Marine Lab, 135 Duke University Marine Lab Road, Beaufort, NC 28516 USA
2-Institute of Ecology, University of Georgia, Athens, GA 30602 USA
3-Department of Biology, Wesleyan University Middletown, CT 06459 USA

Nesting beach studies continue to be the lynch pin of sea turtle population assessment and trend analysis. They are important socially and politically as well, being visible and accessible points of entry for community and stakeholder involvement.  Classical techniques that assume independence of observation are not appropriate for such scenarios and can lead to biased estimates. Mixed models (incorporating fixed and random effects) are useful for modeling observations more akin to real world situations where data are noisy, unbalanced and composed of multiple error terms. We examine the potential of mixed models to provide precision and statistical power comparisons of survey protocols with differing temporal coverage.  Knowing the temporal variability in beach attendance is crucial for calculating power and making design decisions.  We used mixed models to calculate the size of these components of variability (e.g. day to day, year to year and a year-day interaction) and evaluate how changes to the sampling design affect the ability to detect trends in population numbers. We used data from 17 years of saturation tagging of hawksbills (Eretmochelys imbricata) at Pasture Bay, Long Island Antigua to test the statistical power of monitoring designs involving shorter survey periods than the current design of 155 nights each year. There has been virtually no use of mixed models in sea turtle research. Given their appropriateness for the kinds of data generated in recovery and conservation planning for sea turtles mixed models provide a novel approach to answering sea turtle conservation and management questions.