Risk-based framework for the selection of socially optimal dam safety goals

The design, erection, and operation of dams, unlike the vast majority of other engineered systems, is one human activity that has potential for catastrophic destruction. The safety and performance adequacy of the thousands of existing and future dams is an important matter of public policy and welfare. This dissertation investigates, focusing on hydrologic design criteria, the possibility of improving the methods used for the selection of dam safety goals. A review of current practice in the United States reveals a generalized consensus on using conservative criteria (plausible worst case scenarios) based on deterministic models. This strategy has led dam designers to define safety in the narrow sense of an equivalence to low probability of structural failure. Two serious problems encountered in the use of this approach are an unwarranted sense of security concerning flood-control works, and the misallocation of federal resources dedicated to the mitigation of natural hazards. An alternative set of criteria and methods for the selection of dam safety goals is presented. A central premise of the proposal is that the prevention of property damage and loss of lives are different problems that have to be treated separately. The suggested risk-based framework puts the determination of dam safety goals on a more scientific and sensible basis. Potential benefits include the opportunity to achieve a more efficient allocation of federal resources, reduce the likelihood of unanticipated risks from dam malfunctions, devise more effective strategies to minimize loss of life due to flooding, and enhance the safety of dams in the United States and other countries.