Coupling data-driven agent-based and hydrological modelling to explore the effect of collective water allocation strategies in environmental flows

Exploring the human dimension in water resource decision-making has become increasingly relevant for enhancing the effectiveness of water policies, notably water allocation strategies. The complexity of humanwater system interactions poses a challenge to the representation of the dynamics of the system and its regulation. This study presents a framework for socio-hydrological analysis grounded on a data-driven Agent-Based Model (ABM) coupled to a calibrated hydrologic model on WEAP platform to analyse water allocation strategies. A probabilistic methodology was introduced to reproduce the variability of human behaviour. A two-way coupling mechanism between the two models is proposed and enhances the modelled system representation compared to the unidirectional coupling. This framework was applied to a river basin in the Brazilian Cerrado where conflicts among water users have jeopardized water security, agriculture activities and environmental flows. Farmers were classified according to their willingness to cooperate profiles, corresponding to different withdrawal patterns and compliance with water regulations. Mechanisms of collective allocation of water grants represented by a total adherence to each profile and neighbourhood effect were simulated. Scenarios with a higher incidence of non-cooperative agents resulted in flows that are closer to the minimum reference flow. In contrast, scenarios involving collective cooperative decisions led to greater simulated water availability, thus highlighting the importance of cooperation regarding compliance with water use rules to the maintenance of flows within safe levels. The modelling framework can facilitate collective water allocation strategies through the representation of water users' impact in the system.