Concerns stakeholders bring forth during the problem formulation stages of an investigation are often difficult to link with a physical system simulation. We use open ended elicitation, narrative analysis, and stakeholder focus sessions to develop a value model for a water allocation decision problem. Open-ended stakeholder elicitation allows the articulation of non-monetary value-based narratives that accurately reflect the range of values in a problem and “make better use of” value information. Loosely guided elicitation was sequenced in a traditional narrative style to identify stakeholder interpretations of key events and variables relevant to a test case. The concerns stated by stakeholders during interviews are matched to natural attributes of a simulated groundwater system to obtain a method for optimizing policy alternatives using 6 objective functions and an over-arching scientific goal. Attributes that determine the level for objectives serve as the transfer points between community preferences and physical system behavior. A linked simulation-optimization decision support system (DSS) is used to generate a set of alternatives for consideration within the value model. Hypothetical preference sets are applied to eliminate candidate alternatives and evaluate the question of whether the explicit formulation of an aggregate multi-objective function is needed for policy relevant discourse.

This methodology is applied to the Barton Springs segment of the Edwards Aquifer with a groundwater DSS that includes MODFLOW, system dynamics, and metaheuristic searches. a state-approved Groundwater Availability Model is used, the Groundwater Conservation District sets pumping regimes, and stakeholders establish future land-use conditions. We identify zones where pumping could be more concentrated and yet achieve stakeholder concerns. Key hydrogeologic variables for the system are recharge and drought trigger levels. The DSS reveals that some highly ranked alternatives may meet the preferences of more than one stakeholder and achieve similar quantitative aquifer performance.