The spatial distribution of Antarctic biological diversity is needed to provide a basis for more robust conservation.
In Antarctica, data-driven bioregionalisation is an emerging, but still developing field. We therefore need to re-examine its processes and inherent uncertainties to ensure the classifications are valid. Using an existing RSR bioregionalisation developed by LCR with the Gradient Forests algorithm, we will explore the inherent uncertainty in bioregionalisation (algorithm and clustering processes), and its sensitivity to certain types of data (i.e. does climate, soil, or water play a greater role in the spatial distribution of biological diversity of the region). We will provide a forward-looking bioregionalisation based on the existing model that can account for, and visualise potential changes in distributions caused by changes in climate, water availability and the landscape (e.g. increased ice-free areas through glacial retreat). Policy makers will be able to understand how changing environmental data will affect the size, creation, or even the existence of the current bioregions. These analyses will reveal mathematical relationships between environmental covariates and the distribution of species that will provide new insights into Antarctic biogeography and probability of new incursions of non-native species.