The latitudinal diversity gradient is longest known recognized pattern in ecology. We present the first global-scale study of the importance of energy and water availability, water amount, habitat heterogeneity, and historical factors in determining broad-scale patterns of rarity-weighted richness (RWR). Many studies have provided robust inferences that the use RWR has become a valuable tool to conserve least concern and imperiled species. They all found that RWR was more effective than richness-based greedy algorithms in representing species diversity and can provide near-optimum solutions for reserve selection problems. In the last 40 years, several studies have explored how historical factors, topography, climate, climate variability, biotic processes, human actions and disturbance affect species richness. This study jumps start the study of environmental drivers of RWR, such that we can rapidly cover many of the issues that biogeographers addressed in 40 years of research on the environmental drivers of species richness. Understanding the determinants of RWR constitutes one of the key challenges in ecology. The goals of this paper are to (1) produce global maps of RWR of amphibians, birds, and mammals; (2) to identify the environmental, human and historical predictors of RWR for these vertebrates; (3) to describe the functional relationships between predictors and RWR. Maps for amphibians, mammals and birds were analyzed to describe the biogeographical distribution of RWR values for vertebrates. A set of 8 environmental, historic and human predictor variables along with random forests models were used to investigate the association between predictors and RWR.
Results/Conclusions
The geographical patterns of amphibian, bird, and mammal RWR were similar. RWR values increased from the poles to the tropics. A high concentration of RWR was observed in tropical and subtropical forests of the Neotropical, Afrotropical and Indo-Malay biogeographical realms. Random forests models explained 67%, on average, of the variance in RWR patterns (range 55-81%). The variance importance measures, as produced by RF models, indicated that topography, as measures by elevation range, contributed the most to explain the variation in RWR patterns for all groups. We also observed that energy, represented by mean annual temperature, and history were the second most important for RWR variation for amphibians, mammals and birds, respectively. We conclude that the topographic hypothesis, originally developed for plant diversity gradients, offers a parsimonious explanation for vertebrate RWR patterns as well.