Continental scale patterns and predictors of fern richness and phylogenetic diversity

Because ferns have a wide range of habitat preferences and are widely distributed, they are an ideal group for understanding how diversity is distributed. Here we examine fern diversity on a broad-scale using standard and corrected richness measures as well as phylogenetic indices; in addition we determine the environmental predictors of each diversity metric. Using the combined records of Australian herbaria, a dataset of over 60,000 records was obtained for 89 genera to infer richness. A molecular phylogeny of all the genera was constructed and combined with the herbarium records to obtain phylogenetic diversity patterns. A hotspot of both taxic and phylogenetic diversity occurs in the Wet Tropics of northeastern Australia. Although considerable diversity is distributed along the eastern coast, some important regions of diversity are identified only after sample-standardization of richness and through the phylogenetic metric. Of all of the metrics, annual precipitation was identified as the most explanatory variable, in part, in agreement with global and regional fern studies. However, precipitation was combined with a different variable for each different metric. For corrected richness, precipitation was combined with temperature seasonality, while correlation of phylogenetic diversity to precipitation plus radiation indicated support for the species-energy hypothesis. Significantly high and significantly low phylogenetic diversity were found in geographically separate areas. These separate areas correlated with different climatic conditions such as seasonality in precipitation. The phylogenetic metrics identified additional areas of significant diversity, some of which have not been revealed using traditional taxonomic analyses, suggesting that different ecological and evolutionary processes have operated over the continent. Our study demonstrates that it is possible and vital to incorporate evolutionary metrics when inferring biodiversity hotspots from large compilations of data.