A probabilistic approach to niche-based community models for spatial forecasts of assemblage properties and their uncertainties


Correspondence: Loïc Pellissier, Department of Ecology and Evolution, University of Lausanne, Bâtiment Biophore, CH-1015 Lausanne, Switzerland.

E-mail: loic.pellissier@unil.ch



Conservation strategies need predictions that capture spatial community composition and structure. Currently, the methods used to generate these predictive maps generally focus on deterministic processes and omit stochasticity and other uncertainty in model outputs. Here we present a novel approach to model the means and variance of assemblage properties.


The western Swiss Alps.


We propose a new approach to processing probabilistic predictions derived from stacked species distribution models (S-SDMs) in order to predict and assess the uncertainty in predictions of community properties. We compare the utility of our novel approach with that of a traditional threshold-based approach. We used data sampled in 2009 and 2010 from 192 sites in total for mountain butterfly communities spanning a large elevational gradient as a case study and evaluated the ability of our approach to model the species richness and phylogenetic diversity of communities within an ensemble forecasting framework.


Our approach allowed mapping of the variability in species richness and phylogenetic diversity projections, in addition to the mean, for 78 butterfly species. S-SDMs reproduced the observed decrease in phylogenetic diversity and species richness with elevation, a consequence of environmental filtering. The prediction accuracy of community properties varied along environmental gradients: at low elevations, variability was higher for predictions of species richness while it was the opposite for phylogenetic diversity.

Main conclusions

The use of our probabilistic approach to process species distribution model outputs in order to reconstruct communities provides an improved picture of the range of possible assemblage realizations under similar environmental conditions given modelling uncertainty, and helps to inform managers of the usefulness of modelling results.