An Agent-Based Model to evaluate the performance of water allocation methodologies in coupled human-water irrigation systems: the case of the Formoso River Basin (Brazil).

The interface of water systems with political, social and economic contexts adds complexity to water allocation issues, especially in river basins characterized by water access conflicts. Traditional modeling approaches have failed to consider social interactions and impacts in the representation of hydrological processes when considering human actions as fixed boundary conditions. As an alternative, agent-based models (ABMs) have demonstrated flexibility to represent the complexity of individuals or institutions behavior in the modeling of water resources systems. This work presents an ABM for the Formoso River Basin (FRB) in central region of Brazil. The FRB model uses a new cognitive agent architecture based on the BDI (Belief-Desire-Intention) (Bratman, 1987) paradigm integrated into the GAMA modeling platform (Taillandier et al., 2016).

The FRB is characterized by an intense agricultural activity as it provides water for the second largest public irrigation project in Brazil. The high level of irrigation water demand combined with limited water availability led to a critical water shortage event in 2016 resulting in temporary suspension of water allocation in the region. As a result, conflicts among water users were intensified.

The production of soybean seeds during the dry season (May-Sep) with the practice of subsurface irrigation defines the region’s critical period in terms of water scarcity. By coupling hydrological scenarios with an ABM to represent both farmers and water regulators as BDI agents in the FRB socio-hydrological system, our model aims to provide a framework to illustrate the performance of different management strategies, such as negotiable water allocation, collective/private water allocation, water allocation based on water level rules. These alternatives were evaluated in terms of their effectiveness to reduce the conflicts among water users (farmers) in the FRB. The reasoning of farmers decision was based in three main behavioral profiles: 1) cooperative; 2) undefined; 3) resistant. The water regulators were represented in two main behavioral profiles: 1) active; 2) ineffective. The performance of the management strategy was analyzed using economic (farmers income), ecological (water level) and social (farmers equity) performance metrics.

We believe that this work may contribute to important discussions related to the 18^th UPH.