Sustainability and comanagement of subsistence hunting in an indigenous reserve in Guyana

Although hunting is a key component of subsistence strategies of many Amazonians, it is also one of the greatest threats to wildlife. Because indigenous reserves comprise over 20% of Amazonia, effective conservation often requires that conservation professionals work closely with indigenous groups to manage resource use. We used hunter‐generated harvesting data in spatially explicit biodemographic models to assess the sustainability of subsistence hunting of indigenous Waiwai in Guyana. We collected data through a hunter self‐monitoring program, systematic follows of hunters, and semistructured interviews. We used these data to predict future densities of 2 indicator species, spider monkeys (Ateles paniscus) and bearded sakis (Chiropotes sagulatus), under different scenarios of human population expansion and changing hunting technology. We used encounter rates from transect surveys and hunter catch‐per‐unit effort (CPUE) to validate model predictions. Paca (Cuniculus paca) (198 /year), Currosaw (Crax alector) (168), and spider monkey (117) were the most frequently harvested species. Predicted densities of spider monkeys were statistically indistinguishable from empirically derived transect data (Kolmogorov–Smirnov D = 0.67, p = 0.759) and CPUE (D = 0.32, p = 1.000), demonstrating the robustness of model predictions. Ateles paniscus and C. sagulatus were predicted to be extirpated from <13% of the Waiwai reserve in 20 years, even under the most intensive hunting scenarios. Our results suggest Waiwai hunting is currently sustainable, primarily due to their low population density and use of bow and arrow. Continual monitoring is necessary, however, particularly if human population increases are accompanied by a switch to shotgun‐only hunting. We suggest that hunter self‐monitoring and biodemographic modeling can be used effectively in a comanagement approach in which indigenous parabiologists continuously provide hunting data that is then used to update model parameters and validate model predictions.