The value of expert-driven niche models and historical records in informing the sustainable harvest of yellow anaconda in northern Argentina

Background/Question/Methods

Risk of global and local extinction disproportionately impact species with narrow climatic and ecological requirements, as well as limited dispersal ability. As such, reptiles are under particular threat due to their high physiological sensitivities, restricted range size, and habitat-specific adaptations. Some species also suffer intensive extraction for a variety of uses, chiefly the food and skin trades. The yellow anaconda (Eunectes notaeus) is distributed throughout the Paraguay River drainage, from the Pantanal Region in Bolivia, Paraguay, and Brazil, to northeastern Argentina where it reaches the southernmost part of its range. E. notaeus is widely traded for its skin and a sustainable management program endorsed by the Government of Argentina has been in place for the last decade. Using occurrence records from the literature, field observations, and museum specimens, we employ an expert-driven maximum-entropy model to characterize the ecological niche of E. notaeus. We focus on the identification of barriers to dispersal which may isolate sub-populations and thus increase sensitivity to the disturbances imposed by the harvest program. In addition, we quantitatively assess the efficacy of each data type (literature, field observations, historical records) in predicting the estimated species range.

Results/Conclusions

Nine of the original 25 environmental variables were included in the final model following nine iterations of the expert-driven variable selection procedure. This extensive expert collaboration was invaluable to developing a model with high biological realism and increased accuracy, essential for effective conservation decision making. A persistent barrier between northern and southern populations in Argentina was identified and attributed to the presence of a dry, sandy soil-type; a substrate representing inhospitable habitat for the wetland-dependent species. This dispersal barrier is consistent with previous findings from genetic studies indicating localized isolation of subpopulations within the region. This finding has direct implications for the sustainable management of E. notaeus, as some sub-populations may be disproportionately vulnerable to harvesting. Overlay analyses clearly demonstrate that expert-validated historical and observation records are comparable to published data in the extent to which they are capable of predicting the estimated species range. Integration of occurrence data from multiple sources represents a valuable undertaking in ecological niche modeling by extending the sample size and geographic range of the available data, which is of particular utility for lesser known species such as the yellow anaconda.