A goal programming approach to evaluate agroforestry systems in Eastern Panama.

Agroforestry is hypothesised to increase ecological and economic functions of farms. Yet it is unclear if and how much agroforestry should be embedded in diversified farming systems to satisfy farmers' needs while potentially enhancing environmental services. To address this research gap we use a mathematical programming model to investigate the role of different agroforestry systems in hypothetical farm portfolios that reduce trade-offs between farmers' goals. Our approach is innovative because it simultaneously considers multiple objectives and the effect of land-use diversification within a farm, is based on knowledge and perceptions of local farmers, and accounts for heterogeneity in farmer judgement. We test the model in a forest frontier region in Eastern Panama, using data from farmer interviews. Farmers evaluated conventional land uses and two agroforestry systems (silvopasture and alley cropping) against 10 pre-defined socio-economic and ecological objectives. First we determined the optimal farm land-use composition that reduces trade-offs between the 10 objectives. The model selects the mix of land uses that secures the best worst-case performance across all objectives, when considering uncertainty in the ability of each land use to achieve each objective (which we quantify by the variability in farmer opinion). Agroforestry dominates the optimised farm portfolio, which comprises 60% silvopasture, 39% forest and 1% plantation. This land-use portfolio, however, deviates strongly from the current land use of farmers, which is 59% pasture, 26% crops, 14% forest and 1% plantation. In a second step we explore the implicit objectives driving farmers' current land-use decisions. We find that immediate-term needs related to food security and liquidity best explain farmers' current land-use portfolio; optimising for these objectives produces a land-use portfolio comprising 60% pasture and 40% crops, which is similar to the current land use. This suggests that increasing agroforestry adoption in the study area will require systems that provide early and frequent returns and allow for ongoing crop production, to better satisfy farmers' cash flow and household consumption needs.