Revisiting decision-making assumptions to improve deforestation predictions: Evidence from the Amazon

Commodity agriculture is one of the primary drivers of global deforestation, although the contribution of smallscale agriculture is increasing. Understanding deforestation requires comprehension of the human decisionmaking processes that drive land-use choices. Despite that, there are limited studies about the decisionmaking process of non-Western Educated Industrialized Rich and Democratic societies. Hence, research and policies on land use/land cover change often assume smallholders' behavior is driven by monetary/food goals (Income Optimization), disregarding previous evidence suggesting otherwise. People may seek to minimize work drudgeries (Time Optimization) or may establish a minimum amount of working time (Time Budget). Through an agent-based model, we investigated which decision-making assumptions-Time Optimization, Time Budget or the combination of both- best explained Kh & itilde;setje's behavior, a Brazilian Amazon indigenous society, by comparing deforestation predictions with historical records. Our results suggest indigenous people's decisions are better represented by the (less used) minimum working time assumptions (Budget rule). Models following the Budget rule were also less sensitive to unpredictability in the food amount obtained by a household. The results imply that time categorization may be more prevalent than initially anticipated and could contribute to managing the unpredictability of natural resource availability.