Selecting representative reserve systems in remote areas: A case study in the Nullarbor region, Australia

Abstract

We present a procedure for the selection of a network of reserves representing the biological diversity of a large and biologically poorly known region. The quantitative analysis gives equal weighting to a wide array of different species: mobile and sessile, long-lived and ephemeral, heterothermic and homeothermic, etc. Sampling was based on quadrats that were positioned through the region using a stratified random strategy. This provided a presence-absence matrix of the species composition at each quadrat. Numerical pattern analysis was used to identify 14 species assemblages and to re-order the sites and species in the data matrix. The geographic pattern of each assemblage throughout the region was described by contouring assemblage richness, the isolines being the proportion of the number of species in each assemblage. The data matrix was re-examined. Some assemblages exhibit several gradients in species composition. For example, a 0·1 isoline in the east may represent a different 10% of species than a 0·1 isoline in the west. Other assemblages exhibit only a single gradient in species composition. The results were used to select the optimum positions of reserves needed to represent the compositional diversity of each of the 14 species assemblages. Limited field checking confirmed the predicted isolines in assemblage compositional richness to a satisfactory extent.