Designing crop rotations to sustain territorial food self-sufficiency

Ensuring food security in the context of agroecological transitions and environmental challenges requires rethinking agricultural land use at the territorial scale, particularly in a context of increasing global supply chain disruptions. This study develops a mixed-integer linear programming (MILP) model to design agronomically feasible crop rotations that align regional agricultural production with the nutritional needs of populations following healthy and sustainable dietary frameworks. Using Wallonia (Belgium) as a case study, the model integrates pedo-climatic heterogeneity, crop rotation constraints (return times and exclusion constraints), and regional yield potentials to generate spatially explicit rotation-based land allocation strategies.The model was applied under four optimization configurations with progressively increasing levels of agronomic constraints and two dietary frameworks (TYFA and EAT-Lancet). The progressive incorporation of these constraints improved the agronomic coherence of the optimized crop rotations. Under the TYFA scenario including the full set of constraints, territorial food self-sufficiency was achieved, covering 103 % of food demand while using 90 % of available arable land. Under the EAT-Lancet assumptions, food self-sufficiency was also achieved but required a higher land mobilization (99 %) and generated larger production surpluses.Introducing agronomic constraints revealed a clear trade-off between land-use efficiency and agronomic feasibility, reinforcing the prioritization of the most productive regions, while less productive areas primarily acted as adjustment variables. The analysis of crop diversity typology further reveals that the optimized system converges toward rotations of medium-to-high structural and functional diversity, dominated by balanced combinations of cereals, legumes, and oilseeds, with occasional longer and more complex sequences enhancing agronomic robustness. Legumes emerged as the main limiting crop due to the combination of long return time and high demand in plant-based diets. Beyond demonstrating the feasibility of territorial food self-sufficiency, the framework provides a transferable decision-support tool to support the design of resilient territorial food systems.