Estratégias e áreas prioritárias à conservação de quelônios aquáticos e semi-aquáticos na Amazônia

Abstract

This study aimed to evaluate some threats to freshwater turtles in Brazilian Amazon in order to propose strategies and priority areas to their conservation. In the first chapter, we used chelonian distribution data to generate species distribution model (SDM) that predict suitable areas for the occurrence of the group and to evaluate if turtles are protected by the current reserve networks (Gap Analysis). We identified only one gap species, Mesoclemmys nasuta. Other chelonian species had at least a portion of their distribution range included in protected areas. Considering Integral protected Areas (IPA), only Rhinoclemmys punctularia and Kinosternon scorpioides are protected. The insertion of Sustainable Use Areas (SUA) and Indigenous Territories Areas (ITA) was crucial to consider protected most of turtles. Chelonians may be overharvested in non-integral protected areas due their importance as a food resource. It is necessary to restructure the protected areas in Amazon in order to contemplate river catchment sites. In the second chapter we evaluated the exposition of the freshwater turtles to deforestation. We also identified priority areas to the group conservation based on the habitat requirements of the species, deforestation and the network of protected areas. The priority areas to freshwater chelonian conservation are located in northern Amazon and they do not encompass high deforestation areas. However, higher vulnerability of turtle richness to deforestation is located in central-northeastern Amazon, where the deforestation is currently progressing. Phrynops geoffroanus, Kinosternon scorpioides and Rhinoclemmys punctularia had higher percentage of their SDMs lost due to deforestation. Protected areas included large amount of sites required by chelonian conservation but they are not located in areas with higher turtle richness. Our results highlight the regions more important to invest in conservation of freshwater turtles in Brazilian Amazon. In the third chapter, we analized if the Ecological Niche Model (ENM) of a large distributed species, Podocnemis expansa, are transferable to other parts of its geographical range. All TSS values were low, indicating that the models cannot be transferred from one Amazon area to another. On the other hand, AUC showed excellent and good values, no matter the dataset and predictor selection. The predicted area in number of pixels showed strong relationship with the TSS and AUC values using both predictor selections. Our study demonstrates the importance of capture relevant drivers of change in the requirements of the species in each region. Given the lack of transferability, the sum of the suitable areas predicted to both regions of the dataset which produced the more accurate model may be used to represent species distribution. In the fourth chapter, we determined sex ratio of P. unifilis hatchlings based on the dependency of the embryo growth rate from a time series of incubation temperatures of natural nests in central Amazonia. Our study is the first one to show that temperature influences the embryo developmental rate and sex determination in a different way. Daily mean nest temperature was related to distance to vegetation, air temperature and area. Daily thermal variance is related to daily mean nest temperature and both are responsible to sex determination. The sand beach shows high feminization pattern, but the flooding site is still an important source of male production and it seems to be crucial to the maintenance of the population viability.