Revealing areas of high nature conservation importance in a seasonally dry tropical forest in Brazil: Combination of modelled plant diversity hot spots and threat patterns

Abstract

The Caatinga biome has been identified as one of the important wilderness areas on earth. However, less than 1% of the region is under strictly legal protection although Seasonally Dry Tropical Forests (SDTFs) are globally highly endangered. There is an urgent need to increase the understanding of diversity pattern and threaten status of Caatinga plant species to preserve the unique biodiversity and protect endangered species. Species distribution modelling (SDM) can support strategic decisions in nature conservation for poorly researched tropical regions. This study provides the first highly representative, spatially explicit overview of plant diversity and threat status for the entire Caatinga, a semi-arid area in Northeast Brazil. For this purpose, we developed (a) a stacked species distribution modelling (S-SDM) approach to predict quantitatively floristic species richness and patterns of threatened plant species and (b) a combined approach of diversity hot spots and hubs of threatened species to derive conservation importance units (CIU) to contribute to improved nature reserve management. We applied the modelling technique MaxEnt to establish predictive distribution models with 22 uncorrelated predictors including climate, topography, solar radiation and soil information at a high spatial resolution of 30 arc-seconds (approx. 1km). Spatial patterns of species richness and threat status were derived by stacking 1062 plant species and 27 endangered species, respectively. These outputs were compared to two levels of protected areas (Brazilian and international standards) and intensive human land use patterns to define a set of recommendations for conservation management. Our complementary S-SDM approach showed that our predicted CIUs covered an area of 24% across the entire biome, whereas only 7% of the Caatinga is currently protected based on the Brazilian standards. We found that apart from an excellent overlap of 38% between CIUs and the current protected areas, a substantial proportion of CIUs (89%) was predicted outside the existing reserve network. Moreover, our findings enable targeted surveys to be done in order to enhance conservation efforts and ensure the efficient use of available resources in this poorly studied tropical region. Future upcoming local and regional studies could focus on a multi-taxonomic approach including e.g. insects, reptiles or mammals as a holistic perspective towards biodiversity conservation.