Résumé
Crop losses caused by pests and diseases decrease the incomes and threaten the livelihoods of thousands of families worldwide. A good example of the magnitude of these impacts are the massive crop losses experienced by coffee farmers in Central America due to coffee leaf rust. Coffee farmers need agroecosystems that are capable of regulating the negative impacts of pests and diseases while providing other ecosystem services on which their households and society depend. In this study, we aimed to identify the most promising coffee agroforestry systems for regulating diseases and ensuring the provision of other ecosystem services. During two years, in a research network of 61 coffee plots under a wide variety of shade and management conditions in Turrialba, Costa Rica, we quantified primary and secondary coffee losses (yield and economic losses) and indicators of three other ecosystem services: provisioning of agroforestry products (bananas, plantains, other fruits, and timber), maintenance of soil fertility and carbon sequestration. We then performed an analysis of the relationships between losses and ecosystem service indicators. Based on the results of relationships and on three criteria, we identified the coffee agroforestry systems that had the lowest losses due to diseases and that provided desirable levels of agroforestry products, soil fertility and carbon sequestration. We found multiple significant relationships between losses and ecosystem services (including both tradeoffs and synergies) which allowed us to derive recommendations for better management strategies to reduce yield losses. We identified six coffee agroforestry systems (CAFs) as the most promising ones for reducing losses while simultaneously providing other ecosystem services. One of these CAFs was a simple agroforestry system (dominated by service trees), three were medium diversified CAFs and two were highly diversified CAFs (systems including service trees, timber trees, fruit trees and musaceas). The six CAFS differed in their cropping practices and farmer profitability objectives. The six CAFs offer several options for the design of new coffee plantations or for the transformation of existing plantations. Several of this promising CAFs use little fungicides, which is an indicator that the reduction of chemical inputs could be possible. Our results suggest that the regulation of diseases and associated losses in agroforestry systems should be based on, and take advantage of, the positive effects of plant biodiversity, adequate shade cover, good soil fertility, and minimal use of fungicides.