Reserve selection for conserving groundwater biodiversity

1. A possible conservation strategy to minimise the risk of groundwater biodiversity loss due to human activities consists in designing a network of reserve areas at the continental scale that collectively include most groundwater species. To this end, we compared the efficiency of three area selection methods (species richness hotspots, endemism hotspots and complementarity) and examined the influence of spatial constraints (reduced extent and increased aggregation of reserve areas) on the representation of 1059 groundwater species in six European regions. 2. Presence data from a data base elaborated as part of a European initiative on groundwater biodiversity, the PASCALIS project, were referenced onto 4675 grid cells (0.2 by 0.2°). Complementary performed much better than traditional selection methods for maximising species representation in a reserve network arbitrarily limited to 10% of all the cells containing groundwater fauna. It captured 155 more species than areas selected on richness and 77 more species than areas selected on endemism hotspots. 3. Representing species in a specified proportion of their area of occupancy (i.e. 100%, 50% and 10% of the area of occupancy of species occurring in 1, [less-than or equal to]10, and >10 cells, respectively) required inclusion of 46% of the cells containing groundwater fauna. The reserve network needed to achieve this level of coverage may be too large and fragmented to be implemented and managed in practice. 4. Reduction of the reserve areas to 10% of the landscape containing groundwater fauna and their aggregation into a smaller number of cell clusters resulted in a more realistic reserve network that represented 73.8% (782 species) and 59.1% (274 endemics) of the total number of species and endemics, respectively. 5. We propose several research priorities to improve the design of effective groundwater reserve networks in Europe: (i) devising sampling strategies that reduce uncertainties in the placement of reserves and increase the number of alternative reserve networks and (ii) shifting from a grid-cell selection approach to an aquifer-selection approach that incorporates species representation targets, minimum space requirement and also socio-economic costs related to the vulnerability of aquifers and degree of human activity in the catchment.