The increasing demand of livestock products, larger climate uncertainty and resource degradation, and the rampant poverty of farmers in the semi-arid tropics calls for modernized agroforestry systems (AFS) capable of providing multiple functions. The project will develop regionally specific, trait-based and field tested AFS capable of providing critical agro-ecosystem functions in semi-arid Africa and Central America. The primary objective is to provide farmers with a portfolio of regionally suitable tree and shrub species organized by their traits or attributes, in relation to the provisioning of multiple services, as perceived by the farmers and in terms of fundamental ecological functions. The project will make substantial contributions to AF and ecological science through its integration of theories and concepts from both fields, and will thereby provide a scientifically based model for the design of modernized AFS. The project will identify, and characterize the main factors influencing the adoption/non adoption of AF in selected target areas. It will improve the performance of AFS in dry and marginal areas with particular focus on how species can be assembled according to their traits to improve the multi-functionality of AFS and will identify the potential of new multipurpose tree species suitable for AF in dry and marginal areas by grouping these species according to their functional traits and their capacity to provide critical AFS services. The project will specifically enhance the synergies between the species traits and model how trait combinations contribute to the multi-functionality of AFS including soil, trees, crop/pasture and livestock. We will evaluate the short and medium environmental impacts of AF and its socio-economic implications through sophisticated interdisciplinary models and will produce policy recommendations aimed at promoting AFS and related husbandry practices best adapted to specific local needs, yet universal in scope.