On modeling the interdependency among adaptive reservoir operation, floodplain land-use, and agricultural production: a socio-hydrological approach.

Operating a multi-purpose reservoir is challenging because it requires finding the right balance between competing objectives (flood control and water supply) and because of its potential to affect many social groups beyond one area that are varyingly affected by outcomes of those objectives. For example, some groups (e.g., residents living on floodplains) can be sensitive to floods, while others (e.g., farmers) may be more concerned about water shortages. Climate change and increasing frequencies of hydrological extremes further add to the challenge of achieving a fine balance between these competing objectives. There is growing interest in more adaptive forms of reservoir operation for this regard. Forecast-informed reservoir operation (FIRO) is an example of such adaptive approaches that incorporates short-term weather forecasts to flexibly fine-tune reservoir operation to better operate the system under variability. Previous research has examined various ways for optimizing the trade-off between flood risk and water shortage using FIRO-like approaches and hydrological analysis. However, little research has been done to investigate how a transition to adaptive forms of reservoir operation propagates and affects the behaviors of various social groups beyond a reservoir. Whether and how a rule change lead to hidden misfits in the existing protocols of interaction among reservoir operators and downstream social groups and what are potential consequences of such misfits are currently poorly understood. To help fill the research gap, we developed a stylized model of an integrated system of a reservoir and two social groups (floodplain residents and agricultural producers) that are varyingly affected by flood control and water supply outcomes. In the model, the operator adaptively operates the reservoir by synthesizing the demand from different interest groups, pre-defined operation rules, and weather forecasts to minimize the disaster risk to society. Each water user group can petition their demand to the operator and adapt their settlement or agricultural production decisions in response. Under this setup, we tested effects of different behavioral heuristics of reservoir operators on the outcomes faced by the social groups under various hydrological scenarios. We also tested effects of different configurations of social network structure among the actors on the outcomes. Through this effort, we explored potential robustness-fragility tradeoffs and hidden misfits that can occur under FIRO-like regimes of reservoir operation. Our results show that these test variables can induce fundamental changes in the performance of adaptive reservoir operation. Our findings can potentially be expanded to the study of other complex disaster or shared resource management problems involving multiple actors.