Predictive modeling of suitable habitats for threatened marine invertebrates and implications for conservation assessment in Brazil

摘要

Spatial analysis and modeling tools were employed to predict suitable habitat distribution for threatened marine invertebrates and estimate the overlap between highly suitable areas for these species and the Brazilian marine protected areas (MPAs). Records of the occurrence of species were obtained from the collections included in the Ocean Biogeographic Information System (OBIS-Brazil), with additional records culled from the literature. The distribution data of 16 out of 33 threatened species, with at least ten occurrences in the available records, were selected for modeling by Maxent algorithm (Maximum Entropy Modeling) based on environmental variables (temperature, salinity, bathymetry and their derivatives). The resulting maps were filtered with a fixed threshold of 0.5 (to distinguish only the highly suitable areas) and superimposed on MPA digital maps. The algorithm produced reasonable predictions of the species' potential distributions, showing that the patterns predicted by the model are largely consistent with current knowledge of the species. The distribution of the highly suitable areas showed little overlapping with Brazilian MPAs. This study showed how the habitat suitability for threatened species can be assessed using GIS applications and modeling tools.