Reaction-network modeling of restoration pathways in semi-arid agroecosystems

Semi-arid irrigated agroecosystems are shaped by strong feedbacks between vegetation and soil properties such as salinity, organic matter, and nutrient availability. These feedbacks can generate multiple persistent ecosystem configurations and complex restoration pathways. We introduce a reaction-network-based ecosystem modeling framework in which plant guilds and key soil variables are represented as interacting species. Using Chemical Organization Theory, persistent ecosystem configurations are identified as organizations of the reaction network, while restoration processes correspond to transitions between such organizations. Unlike traditional state-and-transition approaches, ecosystem states emerge here directly from the interaction structure of the soil–vegetation system. Applying the framework to empirically grounded soil parameter ranges from the Haouz–Agafay region in Morocco reveals a structured hierarchy of guild assemblages controlled primarily by organic matter and nitrogen availability. The resulting organizational lattice provides a structural map of feasible ecosystem states and restoration pathways. Beyond the specific case study, the framework links plant–soil feedback ecology with reaction network theory and provides a conceptual basis for strategic management of restoration processes in semi-arid agroecosystems. Future work will integrate spatial dynamics and water management to enable quantitative simulation of large-scale restoration scenarios.