Modelling species richness and life form composition in Sahelian Burkina Faso with remote sensing data

Abstract

Species distribution modelling is an important tool in planning conservation strategies, in predicting effects of environmental change and in delineating patterns of species diversity. Until now, the majority of studies dealing with predicted species distributions were conducted at regional scales by using climatic and biomass parameters with low spatial resolution as environmental variables. On the contrary, informed decision-making regarding conservation priorities mainly takes place on a local scale. We used the genetic algorithm for rule-set production (GARP), environmental layers (derivates of high-resolution LANDSAT ETM+ satellite images) and 13,380 geo-referenced records from collection and relevé data to model distributions of vascular plant species in the Sahelian zone of Burkina Faso. For each species, records were divided randomly in a training dataset for model generation and a test dataset to evaluate the modelled distribution maps. Single distributions of 138 highly accurately modelled species were summed up for life forms and total species richness. Highest values of species richness are found in the NW of the study area. The pattern of vascular plant diversity appears to be closely linked to sandy soils and to water availability: watercourses, tiger bush vegetation and dune ridges are more diverse than pediplain. Chamaephytes show a more pronounced preference of dunes over pediplain than other life forms, while geophyte diversity is highest near watercourses. Hemicryptophyte diversity is most evenly distributed. Therophyte diversity is closest to all vascular plant diversity. Phanerophyte diversity is highest in tiger bush vegetation and close to watercourses. Our approach holds a wide application potential to model species distributions and diversity where environmental data are scarce.