No abstract

AimTo determine whether the method used to build distributional maps from raw data influences the representation of two principal macroecological patterns: the latitudinal gradient in species richness and the latitudinal variation in range sizes (Rapoport's rule).Location World-wide. MethodsAll available distribution data from the Global Biodiversity Information Facility (GBIF) for those fish species that are members of orders of fishes with only marine representatives in each order were extracted and cleaned so as to compare four different procedures: point-to-grid (GBIF maps), range maps applying an α-shape [GBIF-extent of occurrence (EOO) maps], the MaxEnt method of species distribution modelling (GBIF-MaxEnt maps) and the MaxEnt method but restricted to the area delimited by the α-shape (GBIF-MaxEnt-restricted maps). ResultsThe location of hotspots and the latitudinal gradient in species richness or range sizes are relatively similar in the four procedures. GBIF-EOO maps and most GBIF-MaxEnt-maps provide overestimations of species richness when compared with those present in a priori well-surveyed cells. GBIF-EOO maps seem to provide more reasonable world macroecological patterns. MaxEnt can erroneously predict the presence of species in environmentally similar cells of another hemisphere or in other regions that lie outside the range of the species. Limiting this overpredictive capacity, as in the case of GBIF-MaxEnt-restricted maps, seems to mimic the frequency of observations derived from a simple point-to-grid procedure, with the utility of this procedure consequently being limited. Main conclusionsIn studies of macroecological patterns at a global scale, the simple α-shape method seems to be a more parsimonious option for extrapolating species distributions from primary data than are distribution models performed indiscriminately and automatically with MaxEnt. GBIF data may be used in macroecological patterns if original data are cleaned, autocorrelation is corrected and species richness figures do not constitute obvious underestimations. Efforts therefore should focus on improving the number and quality of records that can serve as the source of primary data in macroecological studies.

Abstract:We herein present FactorsR, an RWizard application which provides tools for the identification of the most likely causal factors significantly correlated with species richness, and for depicting on a map the species richness predicted by a Support Vector Machine (SVM) model. As a demonstration of FactorsR, we used an assessment using a database incorporating all species of terrestrial carnivores, a total of 249 species, distributed across 12 families. The model performed with SVM explained 91.9% of the variance observed in the species richness of terrestrial carnivores. Species richness was higher in areas with both higher vegetation index and patch index, i.e., containing higher numbers of species whose range distribution is less fragmented. Lower species richness than expected was observed in Chile, Madagascar, Sumatra, Taiwan, and Sulawesi. OPEN ACCESSDiversity 2015, 7 386

Species distribution models (SDMs) are broadly used to predict species distributions from available presence data. However, SDMs results have been criticized for several reasons mainly related to two basic characteristics of most SDMs: 1) general lack of reliable species absence information, 2) the frequent use of an arbitrary geographical extent (GE) or accessible area of the species. These impediments have motivated us to generate a procedure called niche of occurrence (NOO). NOO provides the probable distribution of species (realized niche) relying solely on partial information about presence of species. It operates within a natural geographical extent delimited by available observations and avoids using misleading thresholds to obtain binary presence–absence estimations when the species prevalence is unknown. In this study the main characteristics of NOO are presented, comparing its performance with other recognized and more complex SDMs by using virtual species to avoid the omnipresent error sources of real data sets.

No abstract