Once the objective has been determined, our methodology leads to the selection of that combination of structures, levels of development for different water uses, and operating procedures that will best achieve the objective.

— Arthur Maass (Design of Water Resources Systems, 1962)

The methods of water-system design developed by an interdisciplinary team at Harvard University are best known for their influence on the study of the feasibility of new water projects. However, as the quotation from Professor Maass, the principal author of "Design of Water Resource Systems," shows, the concept of objective-based design can be applied to operating procedures as well.

In this step, the Drought Preparedness Study (DPS) team compares proposed alternatives against the status quo, measuring how well they meet the objectives developed in Step 2 (i.e., Develop Objectives and Metrics for Evaluation). The team will eliminate or redesign alternatives that do not measure up until they are ready to recommend a plan to decision makers.

"The Economic and Environmental Principles and Guidelines for Water and Related Land Resources Implementation Studies" (P&G) list four characteristics of good plans. These characteristics are general enough that they are appropriate for federal or non-federal planning efforts:

• Completeness:  all the elements required to make the plan work are included in the plan.
• Effectiveness:  the alternative addresses the planning objectives.
• Efficiency:  the ratio of plan outputs to inputs is satisfactory.
• Acceptability:  the plan satisfies decision criteria and does not violate planning constraints.

The initial screening of alternatives should emphasize effectiveness and acceptability.  Table 7 illustrates how this initial screening can be accomplished using decision criteria, planning objectives and constraints. The goal of the initial screening is to eliminate some alternatives and develop a ranking of the remaining alternatives. The process of ranking may help in the continuing effort to communicate and clarify objectives and criteria. The initial screening permits the focusing of study resources on the detailed evaluation of the most promising alternatives.  

Each of the alternatives being seriously considered should now be modeled in the shared vision model. In some cases, teams may decide that each alternative should be represented by a separate model (a modification of the status quo model saved with a different file name). In other cases, teams may decide that alternatives can be more effectively represented by internal “switches” in the status quo model which effects the desired change in water management procedures. The models provide the plan performance and outputs required for detailed evaluation. Using the model, the next level of evaluation can be on the basis of performance measures. For example, the model can be used to estimate how much more frequently would a city have to impose curtailment under an alternative than under the status quo. Table 8 provides a more detailed evaluation of alternatives where the status quo is compared to two other alternatives.

 Table 8:  A More Detailed Evaluation of Alternatives

Sometimes an evaluation using just performance measures is enough. If operational changes can be made that benefit many users and hurt none (including the environment), and the value of the benefits clearly outweigh the administrative costs of instituting the changes, then an evaluation of the economic and social effects of each alternative is unnecessary. But what if there is an alternative that benefits some users and hurts others? Or what if an alternative helps everyone but has a significant financial cost? In those cases, an evaluation of the economic, environmental and social effects of the alternatives is the only way to determine which alternative best addresses the goals and decision criteria.

The evaluation of alternatives should lead to tentative recommendations from the DPS team. 

Other factors to consider during alternatives analysis are how to capture the risk and uncertainty inherent in each alternative and incorporate this information into the shared vision model. 

The definitions of the terms risk and uncertainty as they are applied to water resources management have changed a little over time. Risk refers to some negative consequence with an associated probability, even if that probability is difficult to calculate. Risk in water resources management has until recently been defined as the product of the consequence of events multiplied by the probability of the events, that is to say, as an expected value of damages (Guidelines For Risk, 1992). The classic definition of uncertainty involved those unknowns that could not be expressed in probabilistic terms. In flood damage reduction studies, risk which is an expected value of the damage from extreme but rare floods can be compared to annual or present day costs to determine if it would be cost effective to reduce residual flood damages even further by increasing the size of the flood control project. In strategic water supply studies, the “rare, large event” is the drought, and the risk associated with any strategic supply plan is the product of the expected consequences of future droughts times their probability.

But research and experience has shown that people react differently to the risks of a low probability, high consequence event (a 500-year flood, for example) and a high probability, low consequence event (a 2-year flood), even though they may have the same expected value (Guide Book for Risk Perception and Communication, 1993). Thus a more useful definition of risk has come into use that does not multiply damage by probability: risk is the expression of an undesirable consequence in terms of the probability of it happening.

The concept of risk in tactical drought contingency plans has much in common with the risks associated with flood warning systems that are used to minimize damage from floods larger than the design flood. Risks in drought management include: the risk that a very severe drought will cause a catastrophe;and the risk that the drought response plan will be triggered too often (risking reduced effectiveness of public participation in subsequent droughts) or too late (eliminating water savings that would have been possible had the response been initiated sooner). Both of these risks can be assessed using the shared vision model.  No simple quantification, however, will generally be possible because of the various combinations of severities and durations of droughts. Nonetheless, the use of tools such as the Drought Atlas and simulations with the shared vision model can develop a better informed sense of the risk that can be more clearly communicated to decision makers and elected officials.