Use of community characteristics to predict hunting and game harvests in western Amazonian forests.

Wild game harvesting in Amazonia provides rural residents with protein and cash income but can threaten wildlife populations and forest ecosystem functions. As yet, the socioeconomic and environmental drivers that shape hunter livelihoods remain poorly understood. We studied hunting behavior in the Peruvian Amazon through a quantitative characterization of hunters accounting for community and household factors. Data on livelihood activities from a sample of nearly 3800 households in 232 stratified and randomly selected communities were drawn from a survey of the environmental and socioeconomic characteristics of 919 communities. Our double-hurdle model (i.e., 2-stage statistical model that describes whether a household participates in an activity and the amount they participate) separated household game harvesting decisions into 2 parts: first, based on a selection equation that estimated the decision to engage in hunting as a livelihood strategy and, second, based on a truncated lognormal regression equation that estimated total amount of game harvested by households engaged in hunting. We found that 28% of households surveyed reported hunting and that community factors, such as forest cover and distance to the city, drove hunting participation and harvests, although the factors predicting whether a household hunted differed from those that explained game harvests. Household traits, including initial land assets and household head age, were helpful in identifying hunters in communities. Government and nongovernmental organizations should consider socioeconomic and ecological interactions beyond the individual hunter when developing conservation initiatives. Informed targeting of communities in remote areas of Amazonia promises better allocation of scarce resources for wildlife conservation.