Coupling Agent-Based and Hydrological Modeling to Evaluate Water Allocation strategies in irrigated agricultural regions: the case of Formoso River Basin in the Brazilian Cerrado

Transformations in society that emerge from increasingly changing and unpredictable political and economic contexts have resulted in new challenges for water resources issues. In the case of agricultural activities, climate and socio-economic uncertainties aggravate conflicts among farmers and intensify their competition with other uses. The Formoso River Basin (FRB) is an agricultural river basin in the Cerrado biome in Brazil where frequent drought events have jeopardized agricultural activities (especially soybean crops) and led to judicial measures such as the suspension of water rights for irrigation in 2016. The GAN (High-Level Management) Research Project promotes the monitoring of real-time withdrawal from irrigation pumps and their storage in an open-access online platform. Following the socio-hydrology concept, Agent-Based Models (ABM) are a bottom-up decentralized approach in which autonomous agents share the same natural resource and follow behavioral rules to achieve goals. Conflict resolution over water use requires negotiated and collective decision strategies involving cooperation. In this work, a socio-hydrological modeling framework using the ABM approach is proposed. Farmers are cognitive agents who may act according to three identified cooperation profiles, modeled as different compliance levels with water use rules issued by the water regulator agent. The ABM conception included a novel methodology of extracting heterogeneous behavior information obtained from interviews with local experts and data analysis of historical water withdrawal records from the GAN database. The ABM was integrated into a hydrological model using WEAP to represent reciprocal interactions between human and water systems.