Estimation of Condition-Dependent Dispersal Kernel with Simple Bayesian Regression Analysis

Abstract Empirical ornithologists often analyse dispersal distance by histograms separately drawn for categories of individuals (e.g., sexes), and/or by linear models with normal distribution (e.g., ANOVA). However, theoreticians describe dispersal distance by dispersal kernels with various parametric distributions. Therefore, it is a helpful exercise for empiricists to estimate dispersal kernels from field data. As a model case for such an estimation, we analysed dispersal data of the Ryukyu Scops Owls Otus elegans using a Bayesian Weibull regression model. Estimated dispersal kernels showed that males and individuals fledged from late-breeding nests had short natal dispersal distances and that no factors affected breeding dispersal significantly.