Modeling the distribution of a rare Amazonian odonate in relation to future deforestation

The advance of the deforestation frontier in the Amazon forest, the largest tropical forest and one of the richest ecosystems in the world, has threatened several plant and animal species. A lack of good biogeographical information of their distributions and a shortage of basic knowledge on their ecology hinder the proper evaluation of the vulnerability of those species. We used species distribution modeling techniques to fill these gaps and to estimate the vulnerability of a forest-dwelling odonate endemic from the Amazon, Diastatops nigra. We used the MaxEnt algorithm and compared the efficiency of this method in relation to the type of environmental data set (climate-only and climate+hydrographic environmental variables). We also estimated the decrease in extension of occurrence of D. nigra in relation to a recently developed model for future deforestation also produced with the MaxEnt approach. Predicted suitable areas were isolated patches in the central Amazon and many peripheral areas. In general, those areas had stable climates with low seasonality in rainfall. The Amazon deforestation frontier is expanding mainly from the south. The core area of D. nigra distribution is in the central Amazon, so in the short-term projection, the main threat for this species was not the deforestation itself. However, deforestation may extirpate some peripheral populations of this species and increase isolation among those patches of suitable areas. We suggest the use of this model for prioritizing future odonate inventories targeting the other species of the group.