Geospatial analysis to assess distribution patterns and predictive models for endangered plant species to support management decisions: a case study in the Balearic Islands

Abstract

[eng] Species distribution modelling (SDM) has been used to support biodiversity management in recent years. However, the use of SDM at small scales with geolocation systems to obtain high-accuracy location data remains unexplored. In this study, we focused on Euphorbia fontqueriana, an endangered species to Mallorca (western Mediterranean basin), and we aimed to assess the spatial distribution patterns, and to generate a distribution map of the habitat suitability in a wider area. A differential GPS was used to geolocate all the ramets. We used the global Moran’s I index and local Getis Ord-Gi* method to assess the spatial patterns. We pre-selected derived topographic variables that were generated from LiDAR data (elevation, slope, northness and eastness), the connectivity index, normalised difference vegetation index, and soil type, as the environmental (predictor) variables. In addition, we ran the Maxent model using 1603 occurrence locations and seven environmental variables at a resolution of 2 9 2 m in grid size. The population consisted of 1625 ramets that were clustered (global Moran’s I index = 0.161, z score = 16.599, p value \ 0.001) in several hotspots (i.e. areas with high plant density that were surrounded by areas with high plant densities). The Maxent model, which showed a good performance (AUC training score = 0.977), generated a habitat suitability map that displayed zones of high suitability in other areas away from the natural geographical area. Finally, we discuss the usefulness of this study to guide management purposes.