Combining ecotope segmentation and remote sensing data for biotope and species distribution modelling

The design of appropriate biodiversity conservation actions requires an extensive knowledge of biotope and species distributions. Biodiversity monitoring is often a time-consuming task; however, it can be optimised by biotope and species distribution models. In the Lifewatch project, a database combining segmentation in homogeneous landscape units (“ecotopes”) and environmental attributes derived from regularly updated remote sensing data (land cover, topography, potential solar energy,…) and other data sources (climate and edaphic factors) has been designed. Our aim was to assess the usefulness of this database for biotope and species distribution modelling. As a case study, the distributions of a peatbogs (actual and potential) and of a peatbog specialist butterfly (the cranberry fritillary Boloria aquilonaris (Stichel, 1908)) were independently modelled, using the Random Forest algorithm. The agreement between the biotope and species distribution models was assessed. Our map of predictions was compared to a model derived from a more classical grid-based approach. We observed that ecotope segmentation fitted more closely objective limits on the field, thereby improving the efficiency of biodiversity monitoring. The comparison between actual and potential biotopes allowed us to identify potential restoration areas.