Making more out of sparse data: hierarchical modeling of species communities

摘要: Community ecologists and conservation biologists often work with data that are too sparse for achieving reliable inference species-specific approaches. Here we explore the idea of combining models into a single hierarchical model. The community component model seeks shared patterns in how species respond to environmental covariates. We illustrate modeling framework context logistic regression presence–absence data, but similar structure could also be used many other types applications. first use simulated can improve parameterization especially rare species, which would informative alone. then apply real on 500 diatom show it has much greater predictive power than collection independent models. approach roughly one-third distance decay similarity explained by two variables characterizing water quality, typically preferring nutrient-poor waters high pH, common showing more general pattern resource use.