Interactive Scenario Discovery Tool for Decisionmaking Under Deep Uncertainty

View the PRIM tool to see how it identifies key vulnerabilities using a subset of the parameters and simulation results from the Colorado River Basin Study (CRBS).

To use the tool, first select the output variable to define the vulnerability along with the threshold and threshold type. Next, you can select the uncertainties that you wish to consider as part of the definition of a vulnerability and focus the analysis on a smaller range. This version of the tool includes a subset of variables evaluated as part of the CRBS For example, the default setting considers only futures in which the Interim Guidelines are extended past 2026. Next, select the “Run or Reun PRIM Algorithm” box. The algorithm will then return a “peeling trajectory”—a set of boxes that trade off coverage of unacceptable outcome cases with the density of unacceptable outcome cases within the particular box. Box 1, for instance, restricts the range of simulations to only those in which Lee Ferry minimum five-year mean flow is below 1.2 x 107 acre-feet. The box statistics, which are accessible below the figure, indicate that this box has 100-percent coverage—i.e., it includes all unacceptable cases—but at a density of only 62 percent. The user can examine a different box in the “Select PRIM Box” pull-down menu. Selecting Box 4, for example, identifies three uncertainties—Lee Ferry mean flow, Lee Ferry Minimum 10-year Mean Flow, and Lee Ferry Minimum 5-year Mean Flow. The box definition explains 94 percent of the unacceptable outcome cases, and it has a density of 81 percent. This definition of a vulnerability is useful to the Basin States because it concisely describes key conditions that would very likely lead to poor performance. This information can help guide the development of new strategies and can inform the development of a monitoring program for an adaptive strategy. See the Metropolitan Water District of Southern California Case Study for an example. This box is similar to the one identified in the CRBS. Other peeling trajectories and boxes can be found by including or excluding variables.