Labeling Ecological Niche Models

摘要: The ongoing biodiversity crisis is pushing ecologists and conservation biologists to develop models foretell the effects of human-induced transformation natural resources on distribution species, although ecology biogeography still lacks a paradigmatic body theory fully understand drivers patterns. Two decades research ecological niche species distributions have been characterized by technical development discussions plethora methods or algorithms infer predict distributions. Here we suggest metaphorical classification scheme for some most popular based their complexity, interpretability suitability specific applications in biology. Our purpose not compare capacity accurately observed nor criticize how they are commonly used applied studies. Instead, believe that simple can potentially highlight more suited Envelope distance-based grouped into “fish bowl” category, transparency simplicity. Statistical classified as “turbine” models, because hidden complexity general applicability. Finally, machine-learning “vault” high lack fit parameters. We conclude diversity today expected young field, but choice modeling strategy depends study. provide guidelines choosing studies planning climate change mitigation intermediate forecast good balance between interpretability, predictive power robustness model over-fit.