DISSERTATION RESEARCH: Testing the effects of priors on prediction error in Bayesian demographic models

点击翻译按钮获取中文摘要

It is critical that forecasts of populations sizes and extinction risks applied to rare species have high accuracy and low uncertainty. One possible solution to reducing uncertainty in these models is to include information on closely related species in a Bayesian population viability analysis (PVA). This study tests the ability of Bayesian PVA to predict real future extinction risks and population sizes. To do this, the researchers will construct Bayesian models for two rare plant species, Calochortus tiburonensis and Pedicularis furbishiae, Calochortus tiburonensis and Pedicularis furbishiae, using external data from closely related congeners. These two species were originally studied in the 1980s, and the original populations will be re-censused. Estimates of the extinction risks and 2009 population sizes derived from different models will be compared with information on the populations obtained from contemporary censuses. This research provides one of the first tests to determine the appropriateness of using external data in Bayesian PVA. The results will be applied to improving conservation and recovery plans for the rare species studied. The researchers will make the modeling software available online and teach others how to use it at international and regional conservation conferences. Further this research promotes cross-disciplinary interactions and will train both a graduate student and an undergraduate in the use of Bayesian statistical modeling.