Abstract
Tropical dry forests are among the most threatened terrestrial ecosystems worldwide. This study applies robust multi-objective optimization and Pareto frontier analysis to support sustainable land-use planning in dry forest ecosystems, taking as an example the drylands of southern Ecuador. By integrating ecological and socioeconomic indicator bundles, we modeled optimal land-use compositions under uncertainty and compared them to observed allocations derived from GIS, field data, and farmer input. The observed landscape, dominated by silvopasture (57%) and maize (32%), contrasts with the model’s optimal allocation, which prioritizes shaded cocoa (25%) and coffee (23%), reduces silvopasture (15%), and modestly increases maize (37%). The model enhanced a land-use performance index across different levels of considered uncertainty (low: 22–48%; moderate: 10–32%; high: 16–32%), revealing the method’s strength in generating valuable farm-level insights. The Pareto frontier analysis indicated trade-offs between bundles of ecological and economic indicators, mirroring real-world tensions. While observed land use aligns closely with optimized socioeconomic objectives, it underperforms ecologically. Agroforestry emerges as a promising compromise, though incentives and policy support will be key for adoption. Our findings illustrate how robust multi-objective optimization can strengthen intuitive diversification strategies, balance short- and long-term goals, and guide transitions to more resilient land uses. This is especially critical in vulnerable, data-scarce dry ecosystems increasingly affected by environmental and socioeconomic stressors.