Modelling Native and Invasive Woody Species: A Comparison of ENFA and MaxEnt Applied to the Azorean Forest

Abstract

Species distribution models are algorithmic tools that relate the distribution and occurrence of a species to the environmental characteristics of the location from where it has been recorded. Those models, also known as ecological niche models, have emerged as an effective tool in spatial ecology, conservation and land management. The Ecological Niche Factor Analysis (ENFA) is one of the common modelling approaches that are suitable for predicting potential distributions, based on presences only, providing an ecological interpretation based on marginality and specialization. In Maximum Entropy Modelling (MaxEnt) the relative entropy is minimized between the two probability densities defined in the covariate space i.e. estimated from presence data or from landscape. It focuses on relating the environmental conditions of the area where the species is present to the environmental conditions across the area of interest. ENFA has been successfully used in the Azores to model the potential distribution of indigenous and non-indigenous trees. In this paper we use distribution data from one of the most important woody plant invaders in the Azores, Pittosporum undulatum, to compare both modelling approaches. We also test both methods when using the selected environmental variables to predict the distribution in other islands and for other species (Acacia melanoxylon and Morella faya). In general, the two methodologies derived similar predictions. However, our results suggest that the set of environmental variables selected to model the distribution of a species in one particular island will probably have to be adjusted to fit other regions and species.