DECISION PROBLEM FORMULATION FOR GROUNDWATER ALLOCATION PLANNING

The formulation of a decision problem in many respects drives the resultant policy or management recommendation, particularly when simulation-optimization modelling is used. While subject matter experts can create technically sound problem representations, an approach that segregates modeling and computation from the decision makers risks missing key issues. Often a groundwater model gets left 'on the shelf', rather than going into use, because the original decision problem formulation was not useful in a policy or management context.

Integrated Water Resources Management (IWRM) encourages approaches that shift the role of scientific information away from strict prediction to being a catalyst for collaborative learning and management. When modeling 'wicked' problems for decision support, we argue that it is worthwhile placing greater emphasis on developing appropriate questions to answer, rather than on model predictions to answer the questions, as is usually the case. Ancillary benefits include reducing the potential misuse of models and increasing trust in the model outcomes.

Critical issues that must be considered in problem formulation are discussed with reference to experiences for groundwater allocation cases in both Australia and the United States. At a high level the primary objectives of problem framing are 1) assure that the 'right' questions are addressed, that meet the needs and concerns of decision makers 2) assure that it will be possible to obtain scientifically sound and robust results, at the required spatial and temporal scale, within the allocated resources, and 3) carefully phrasing problems mathematically to assure that the design of decision variables and constraints on the feasibility space create a solution space that can be searched with computational efficiency. To achieve all objectives, scientists, modelers, decision makers and managers all have a responsibility. Taking a considered approach to decision problem formulation, accounting for these different objectives and their underlying issues, will help overcome barriers and leverage existing advantages of simulation-optimisation, resulting in greater inclusion of science in groundwater resource management and policy making.