Guaranteeing wildlife populations' viability is crucial to conserve biodiversity as well as safeguard the food security of forest dwellers. Nevertheless, designing ways of ensuring the sustainability of hunting of game species remains complex. Ecological models can guide conservation actions but require accurate information on population parameters, such as reproductive and mortality rates. Quantifying these rates – especially for tropical forest species – is extremely laborious and costly and most models have simplified these aspects. Such sustainability estimations operate under high levels of uncertainty, and although results often point to hunting being unsustainable, many supposedly over-hunted species are still consistently present within hunters’ catches. This fallacy can jeopardize the livelihoods of people closely linked to wildlife by limiting their access to a crucial resource. In this study, we outline a model building process aimed at improving hunting sustainability estimations and at narrowing the gap in trust between modelers and managers.

We worked on two fronts: 1) we aligned the model and its structure with management objectives and 2) we incorporated complex animal population dynamics based on data collected through a participatory approach with hunters. We used the lowland paca (Cuniculus paca) as model species, one of the most valued species by Amazonian people in terms of both consumption and trade. During past fieldwork in two sites in the Peruvian and Brazilian Amazon, hunters have emphasized the need to improve the management of this game species. In response to this demand, we built an agent-based model using the GAMA platform depicting pacas’ population dynamics within the hunting territory of a hypothetical settlement within the Amazon forest. To validate the model, we compared 15 years simulation results with data collected in situ by hunters for 18 years on pacas’ reproductive parameters, which are important for proper game management. Our comparison showed that the model was effective to reproduce empirical data on the species’ reproductive biology.

To improve sustainability estimations we need first of all to improve how game species are represented in our models, as this will dictate how they respond to hunting pressure. Further, if models are to be useful for environmental decision-making, these need to be harmonized with management objectives: their parameters need to reflect system attributes that can be manipulated by managers while outputs need to reflect measurable attributes. Scientists should capitalize on the increasing amount of data available – especially thanks to participation of resource users – and on modelling approaches that can accomodate the complexity and inherent dynamics of game species reproductive biology.