Abstract:
|
The most commonly used measure of plant abundance is percent cover. Different monitoring programs have generated large collections of cover data sets that span many years and/or large spatial extents. The current practice for analyzing plant cover is to use statistical methods that rely on the normal distribution for quantifying uncertainty. This is problematic because plant cover data tend to be left-skewed (J-shaped), right skewed (L-shaped) or U-shaped and, therefore, commonly violate the assumption of normality. Conversely, the beta distribution can provide an ecologically relevant latent variable when modeling plant abundance. A flexible hierarchical framework allows for the observation process to arise from all sorts of field methods, visual estimates of top cover, ordinal cover classes, pin-point, line-point intercept, or line intercept. We also introduce our class of beta models that allow for multiple types of observation errors (measurement and detection errors). We demonstrate how harnessing the beta distribution is preferred for enhancing ecological interpretation with real data.
|