How to model species along ecological gradients?

Florian JANSEN (Greifswald): The shape of species responses along ecological or time gradients are of great theoretical and practical interest. They have important implications for all kinds of numerical community analyses and the chosen response shape determines model attributes like species optima or niche width. Hierarchical logistic regression models (the best model is chosen from a set of predefined models by statistical information criteria) offer a flexible way to efficiently fit species response data. They propose a sound theoretical background for ecological interpretation and can compete against flexible but more generic approaches like Generalised Additive Models (GAM) in terms of statistical inference. I will present extended so called Huisman-Ollf-Fresco (HOF) models, named by the authors of an article from 1993(!). The original set of 5 model types has been extended by two bimodal types to match common ecological expectations. Using modern techniques implemented in the statistical environment R, the hierarchical models can still be seen as an effective tool for univariate species response modelling.