| This research presents a methodology for handling climate uncertainty in infrastructure design analyses. The methodology includes four primary components: climate model projections (CMPs), hydrologic input-data generation techniques (IGTs), reservoir system models (SMs), and output data sets (ODSs). CMPs provide the bounds for hydrologic scenarios; IGTs generate the hydrologic scenarios; and SMs describe the system and represent it as an optimization model. ODSs are generated by running the scenarios, which are generated with the CMPs and IGTs, through the SM.;Key characteristics of this methodology are: 1) the avoidance of scenario probability distributions, 2) the employment of multiple models, and 3) the inclusion of parameter ranges for defining variables. Model components accommodate various climate models and multiple scenarios based on these climate models; additionally, the system model facilitates various levels of complexity. Thousands of future climate scenarios are calculated. Considering all of these scenarios equally allows the results to be interpreted from a robustness perspective. Other performance metrics used in this research are reliability, vulnerability, and net present value.;This methodology is applied as an aid for determining the active reservoir capacity of a potential reservoir. Initially, historical data is used as the CMP, modified autoregressive equations are used as the IGT, and the SM represents a single-reservoir system. Then, the methodology is applied to the Duc Xuyen reservoir in the Mekong Basin in various contexts described by the following system models. First, a basic system model is used to test the effectiveness of two different objective functions and three different levels of future climate variability. Then, the basic approach scenarios are altered to include surprise events. Next, the basic model is expanded to consider within-year storage yield, releases, and targets. Economic components are also accounted for through the determination of the net present value: the allocation system model compares potential benefits from different water uses, while the hydropower system model tries to identify a reservoir operating policy that maximizes the net present value of potential benefits. Lastly, the considerations of when to build the reservoir are discussed.;All of these experiments illustrate the effectiveness of this approach. The inclusion of so many models and scenarios makes it possible to define reservoir size robustness. The additional performance metrics of reliability, vulnerability, and net present value can to help decision makers decide on the most appropriate reservoir capacity. |
